Search Results

Search found 23309 results on 933 pages for 'bit operations'.

Page 86/933 | < Previous Page | 82 83 84 85 86 87 88 89 90 91 92 93  | Next Page >

  • Is there a way of providing a final transform method when chaining operations (like map reduce) in underscore.js?

    - by latentflip
    (Really strugging to title this question, so if anyone has suggestions feel free.) Say I wanted to do an operation like: take an array [1,2,3] multiply each element by 2 (map): [2,4,6] add the elements together (reduce): 12 multiply the result by 10: 120 I can do this pretty cleanly in underscore using chaining, like so: arr = [1,2,3] map = (el) -> 2*el reduce = (s,n) -> s+n out = (r) -> 10*r reduced = _.chain(arr).map(map).reduce(reduce).value() result = out(reduced) However, it would be even nicer if I could chain the 'out' method too, like this: result = _.chain(arr).map(map).reduce(reduce).out(out).value() Now this would be a fairly simple addition to a library like underscore. But my questions are: Does this 'out' method have a name in functional programming? Does this already exist in underscore (tap comes close, but not quite).

    Read the article

  • Is this a good approach to execute a list of operations on a data structure in Python?

    - by Sridhar Iyer
    I have a dictionary of data, the key is the file name and the value is another dictionary of its attribute values. Now I'd like to pass this data structure to various functions, each of which runs some test on the attribute and returns True/False. One approach would be to call each function one by one explicitly from the main code. However I can do something like this: #MYmodule.py class Mymodule: def MYfunc1(self): ... def MYfunc2(self): ... #main.py import Mymodule ... #fill the data structure ... #Now call all the functions in Mymodule one by one for funcs in dir(Mymodule): if funcs[:2]=='MY': result=Mymodule.__dict__.get(funcs)(dataStructure) The advantage of this approach is that implementation of main class needn't change when I add more logic/tests to MYmodule. Is this a good way to solve the problem at hand? Are there better alternatives to this solution?

    Read the article

  • When machine code is generated from a program how does it translates to hardware level operations ??

    - by user553492
    Like if say the instruction is something like 100010101 1010101 01010101 011101010101. Now how is this translating to an actual job of deleting something from memory? Memory consists of actual physical transistors the HOLD data. What causes them to lose that data is some external signal? I want to know how that signal is generated. Like how some binary numbers change the state of a physical transistor. Is there a level beyond machine code that isn't explicitly visible to a programmer? I have heard of microcode that handle code at hardware level, even below assembly language. But still I pretty much don't understand. Thanks!

    Read the article

  • What PHP function(s) can I use to perform operations on non-integer timestamps?

    - by stephenhay
    Disclaimer, I'm not a PHP programmer, so you might find this question trivial. That's why I'm asking you! I've got this kind of timestamp: 2010-05-10T22:00:00 (That's Y-m-d) I would like to subtract, say, 10 days (or months, whatever) from this, and have my result be in the same format, i.e. 2010-04-30T22:00:00. What function(s) do I need to do this in PHP? Note: I'm using this to do a computed field in Drupal. The result will be the date that an e-mail is sent. Bonus question: If 2010-05-10T22:00:00 means "May 10, 2010 at 10pm", is there a timestamp equivalent of "May 10, 2010 (all day)"? Thanks everyone.

    Read the article

  • Are there any javascript string formatting operations similar to the way %s is used in Python?

    - by Phil
    I've been writing a lot of javascript, and when I want to stick a variable in a string, I've been doing it like so: $("#more_info span#author").html("Created by: <a href='/user/" + author + "'>" + author + "</a>"); I feel like it's pretty ugly and a pain to write over and over. In python the %s operator makes this problem easy. Even in C, I can do sprintf (IIRC). Is there anything like that in javascript? (Lots of google'ing yielded nothing.)

    Read the article

  • obiee memory usage

    - by user554629
    Heap memory is a frequent customer topic. Here's the quick refresher, oriented towards AIX, but the principles apply to other unix implementations. 1. 32-bit processes have a maximum addressability of 4GB; usable application heap size of 2-3 GB.  On AIX it is controlled by an environment variable: export LDR_CNTRL=....=MAXDATA=0x080000000   # 2GB ( The leading zero is deliberate, not required )   1a. It is  possible to get 3.25GB  heap size for a 32-bit process using @DSA (Discontiguous Segment Allocation)     export LDR_CNTRL=MAXDATA=0xd0000000@DSA  # 3.25 GB 32-bit only        One side-effect of using AIX segments "c" and "d" is that shared libraries will be loaded privately, and not shared.        If you need the additional heap space, this is worth the trade-off.  This option is frequently used for 32-bit java.   1b. 64-bit processes have no need for the @DSA option. 2. 64-bit processes can double the 32-bit heap size to 4GB using: export LDR_CNTRL=....=MAXDATA=0x100000000  # 1 with 8-zeros    2a. But this setting would place the same memory limitations on obiee as a 32-bit process    2b. The major benefit of 64-bit is to break the binds of 32-bit addressing.  At a minimum, use 8GB export LDR_CNTRL=....=MAXDATA=0x200000000  # 2 with 8-zeros    2c.  Many large customers are providing extra safety to their servers by using 16GB: export LDR_CNTRL=....=MAXDATA=0x400000000  # 4 with 8-zeros There is no performance penalty for providing virtual memory allocations larger than required by the application.  - If the server only uses 2GB of space in 64-bit ... specifying 16GB just provides an upper bound cushion.    When an unexpected user query causes a sudden memory surge, the extra memory keeps the server running. 3.  The next benefit to 64-bit is that you can provide huge thread stack sizes for      strange queries that might otherwise crash the server.      nqsserver uses fast recursive algorithms to traverse complicated control structures.    This means lots of thread space to hold the stack frames.    3a. Stack frames mostly contain register values;  64-bit registers are twice as large as 32-bit          At a minimum you should  quadruple the size of the server stack threads in NQSConfig.INI          when migrating from 32- to 64-bit, to prevent a rogue query from crashing the server.           Allocate more than is normally necessary for safety.    3b. There is no penalty for allocating more stack size than you need ...           it is just virtual memory;   no real resources  are consumed until the extra space is needed.    3c. Increasing thread stack sizes may require the process heap size (MAXDATA) to be increased.          Heap space is used for dynamic memory requests, and for thread stacks.          No performance penalty to run with large heap and thread stack sizes.           In a 32-bit world, this safety would require careful planning to avoid exceeding 2GM usable storage.     3d. Increasing the number of threads also may require additional heap storage.          Most thread stack frames on obiee are allocated when the server is started,          and the real memory usage increases as threads run work. Does 2.8GB sound like a lot of memory for an AIX application server? - I guess it is what you are accustomed to seeing from "grandpa's applications". - One of the primary design goals of obiee is to trade memory for services ( db, query caches, etc) - 2.8GB is still well under the 4GB heap size allocated with MAXDATA=0x100000000 - 2.8GB process size is also possible even on 32-bit Windows applications - It is not unusual to receive a sudden request for 30MB of contiguous storage on obiee.- This is not a memory leak;  eventually the nqsserver storage will stabilize, but it may take days to do so. vmstat is the tool of choice to observe memory usage.  On AIX vmstat will show  something that may be  startling to some people ... that available free memory ( the 2nd column ) is always  trending toward zero ... no available free memory.  Some customers have concluded that "nearly zero memory free" means it is time to upgrade the server with more real memory.   After the upgrade, the server again shows very little free memory available. Should you be concerned about this?   Many customers are !!  Here is what is happening: - AIX filesystems are built on a paging model.   If you read/write a  filesystem block it is paged into memory ( no read/write system calls ) - This filesystem "page" has its own "backing store" on disk, the original filesystem block.   When the system needs the real memory page holding the file block, there is no need to "page out".    The page can be stolen immediately, because the original is still on disk in the filesystem. - The filesystem  pages tend to collect ... every filesystem block that was ever seen since    system boot is available in memory.  If another application needs the file block, it is retrieved with no physical I/O. What happens if the system does need the memory ... to satisfy a 30MB heap request by nqsserver, for example? - Since the filesystem blocks have their own backing store ( not on a paging device )   the kernel can just steal any filesystem block ... on a least-recently-used basis   to satisfy a new real memory request for "computation pages". No cause for alarm.   vmstat is accurately displaying whether all filesystem blocks have been touched, and now reside in memory.   Back to nqsserver:  when should you be worried about its memory footprint? Answer:  Almost never.   Stop monitoring it ... stop fussing over it ... stop trying to optimize it. This is a production application, and nqsserver uses the memory it requires to accomplish the job, based on demand. C'mon ... never worry?   I'm from New York ... worry is what we do best. Ok, here is the metric you should be watching, using vmstat: - Are you paging ... there are several columns of vmstat outputbash-2.04$ vmstat 3 3 System configuration: lcpu=4 mem=4096MB kthr    memory              page              faults        cpu    ----- ------------ ------------------------ ------------ -----------  r  b    avm   fre  re  pi  po  fr   sr  cy  in   sy  cs us sy id wa  0  0 208492  2600   0   0   0   0    0   0  13   45  73  0  0 99  0  0  0 208492  2600   0   0   0   0    0   0   9   12  77  0  0 99  0  0  0 208492  2600   0   0   0   0    0   0   9   40  86  0  0 99  0 avm is the "available free memory" indicator that trends toward zerore   is "re-page".  The kernel steals a real memory page for one process;  immediately repages back to original processpi  "page in".   A process memory page previously paged out, now paged back in because the process needs itpo "page out" A process memory block was paged out, because it was needed by some other process Light paging activity ( re, pi, po ) is not a concern for worry.   Processes get started, need some memory, go away. Sustained paging activity  is cause for concern.   obiee users are having a terrible day if these counters are always changing. Hang on ... if nqsserver needs that memory and I reduce MAXDATA to keep the process under control, won't the nqsserver process crash when the memory is needed? Yes it will.   It means that nqsserver is configured to require too much memory and there are  lots of options to reduce the real memory requirement.  - number of threads  - size of query cache  - size of sort But I need nqsserver to keep running. Real memory is over-committed.    Many things can cause this:- running all application processes on a single server    ... DB server, web servers, WebLogic/WebSphere, sawserver, nqsserver, etc.   You could move some of those to another host machine and communicate over the network  The need for real memory doesn't go away, it's just distributed to other host machines. - AIX LPAR is configured with too little memory.     The AIX admin needs to provide more real memory to the LPAR running obiee. - More memory to this LPAR affects other partitions. Then it's time to visit your friendly IBM rep and buy more memory.

    Read the article

  • GCC, -O2, and bitfields - is this a bug or a feature?

    - by Rooke
    Today I discovered alarming behavior when experimenting with bit fields. For the sake of discussion and simplicity, here's an example program: #include <stdio.h> struct Node { int a:16 __attribute__ ((packed)); int b:16 __attribute__ ((packed)); unsigned int c:27 __attribute__ ((packed)); unsigned int d:3 __attribute__ ((packed)); unsigned int e:2 __attribute__ ((packed)); }; int main (int argc, char *argv[]) { Node n; n.a = 12345; n.b = -23456; n.c = 0x7ffffff; n.d = 0x7; n.e = 0x3; printf("3-bit field cast to int: %d\n",(int)n.d); n.d++; printf("3-bit field cast to int: %d\n",(int)n.d); } The program is purposely causing the 3-bit bit-field to overflow. Here's the (correct) output when compiled using "g++ -O0": 3-bit field cast to int: 7 3-bit field cast to int: 0 Here's the output when compiled using "g++ -O2" (and -O3): 3-bit field cast to int: 7 3-bit field cast to int: 8 Checking the assembly of the latter example, I found this: movl $7, %esi movl $.LC1, %edi xorl %eax, %eax call printf movl $8, %esi movl $.LC1, %edi xorl %eax, %eax call printf xorl %eax, %eax addq $8, %rsp The optimizations have just inserted "8", assuming 7+1=8 when in fact the number overflows and is zero. Fortunately the code I care about doesn't overflow as far as I know, but this situation scares me - is this a known bug, a feature, or is this expected behavior? When can I expect gcc to be right about this? Edit (re: signed/unsigned) : It's being treated as unsigned because it's declared as unsigned. Declaring it as int you get the output (with O0): 3-bit field cast to int: -1 3-bit field cast to int: 0 An even funnier thing happens with -O2 in this case: 3-bit field cast to int: 7 3-bit field cast to int: 8 I admit that attribute is a fishy thing to use; in this case it's a difference in optimization settings I'm concerned about.

    Read the article

  • How John Got 15x Improvement Without Really Trying

    - by rchrd
    The following article was published on a Sun Microsystems website a number of years ago by John Feo. It is still useful and worth preserving. So I'm republishing it here.  How I Got 15x Improvement Without Really Trying John Feo, Sun Microsystems Taking ten "personal" program codes used in scientific and engineering research, the author was able to get from 2 to 15 times performance improvement easily by applying some simple general optimization techniques. Introduction Scientific research based on computer simulation depends on the simulation for advancement. The research can advance only as fast as the computational codes can execute. The codes' efficiency determines both the rate and quality of results. In the same amount of time, a faster program can generate more results and can carry out a more detailed simulation of physical phenomena than a slower program. Highly optimized programs help science advance quickly and insure that monies supporting scientific research are used as effectively as possible. Scientific computer codes divide into three broad categories: ISV, community, and personal. ISV codes are large, mature production codes developed and sold commercially. The codes improve slowly over time both in methods and capabilities, and they are well tuned for most vendor platforms. Since the codes are mature and complex, there are few opportunities to improve their performance solely through code optimization. Improvements of 10% to 15% are typical. Examples of ISV codes are DYNA3D, Gaussian, and Nastran. Community codes are non-commercial production codes used by a particular research field. Generally, they are developed and distributed by a single academic or research institution with assistance from the community. Most users just run the codes, but some develop new methods and extensions that feed back into the general release. The codes are available on most vendor platforms. Since these codes are younger than ISV codes, there are more opportunities to optimize the source code. Improvements of 50% are not unusual. Examples of community codes are AMBER, CHARM, BLAST, and FASTA. Personal codes are those written by single users or small research groups for their own use. These codes are not distributed, but may be passed from professor-to-student or student-to-student over several years. They form the primordial ocean of applications from which community and ISV codes emerge. Government research grants pay for the development of most personal codes. This paper reports on the nature and performance of this class of codes. Over the last year, I have looked at over two dozen personal codes from more than a dozen research institutions. The codes cover a variety of scientific fields, including astronomy, atmospheric sciences, bioinformatics, biology, chemistry, geology, and physics. The sources range from a few hundred lines to more than ten thousand lines, and are written in Fortran, Fortran 90, C, and C++. For the most part, the codes are modular, documented, and written in a clear, straightforward manner. They do not use complex language features, advanced data structures, programming tricks, or libraries. I had little trouble understanding what the codes did or how data structures were used. Most came with a makefile. Surprisingly, only one of the applications is parallel. All developers have access to parallel machines, so availability is not an issue. Several tried to parallelize their applications, but stopped after encountering difficulties. Lack of education and a perception that parallelism is difficult prevented most from trying. I parallelized several of the codes using OpenMP, and did not judge any of the codes as difficult to parallelize. Even more surprising than the lack of parallelism is the inefficiency of the codes. I was able to get large improvements in performance in a matter of a few days applying simple optimization techniques. Table 1 lists ten representative codes [names and affiliation are omitted to preserve anonymity]. Improvements on one processor range from 2x to 15.5x with a simple average of 4.75x. I did not use sophisticated performance tools or drill deep into the program's execution character as one would do when tuning ISV or community codes. Using only a profiler and source line timers, I identified inefficient sections of code and improved their performance by inspection. The changes were at a high level. I am sure there is another factor of 2 or 3 in each code, and more if the codes are parallelized. The study’s results show that personal scientific codes are running many times slower than they should and that the problem is pervasive. Computational scientists are not sloppy programmers; however, few are trained in the art of computer programming or code optimization. I found that most have a working knowledge of some programming language and standard software engineering practices; but they do not know, or think about, how to make their programs run faster. They simply do not know the standard techniques used to make codes run faster. In fact, they do not even perceive that such techniques exist. The case studies described in this paper show that applying simple, well known techniques can significantly increase the performance of personal codes. It is important that the scientific community and the Government agencies that support scientific research find ways to better educate academic scientific programmers. The inefficiency of their codes is so bad that it is retarding both the quality and progress of scientific research. # cacheperformance redundantoperations loopstructures performanceimprovement 1 x x 15.5 2 x 2.8 3 x x 2.5 4 x 2.1 5 x x 2.0 6 x 5.0 7 x 5.8 8 x 6.3 9 2.2 10 x x 3.3 Table 1 — Area of improvement and performance gains of 10 codes The remainder of the paper is organized as follows: sections 2, 3, and 4 discuss the three most common sources of inefficiencies in the codes studied. These are cache performance, redundant operations, and loop structures. Each section includes several examples. The last section summaries the work and suggests a possible solution to the issues raised. Optimizing cache performance Commodity microprocessor systems use caches to increase memory bandwidth and reduce memory latencies. Typical latencies from processor to L1, L2, local, and remote memory are 3, 10, 50, and 200 cycles, respectively. Moreover, bandwidth falls off dramatically as memory distances increase. Programs that do not use cache effectively run many times slower than programs that do. When optimizing for cache, the biggest performance gains are achieved by accessing data in cache order and reusing data to amortize the overhead of cache misses. Secondary considerations are prefetching, associativity, and replacement; however, the understanding and analysis required to optimize for the latter are probably beyond the capabilities of the non-expert. Much can be gained simply by accessing data in the correct order and maximizing data reuse. 6 out of the 10 codes studied here benefited from such high level optimizations. Array Accesses The most important cache optimization is the most basic: accessing Fortran array elements in column order and C array elements in row order. Four of the ten codes—1, 2, 4, and 10—got it wrong. Compilers will restructure nested loops to optimize cache performance, but may not do so if the loop structure is too complex, or the loop body includes conditionals, complex addressing, or function calls. In code 1, the compiler failed to invert a key loop because of complex addressing do I = 0, 1010, delta_x IM = I - delta_x IP = I + delta_x do J = 5, 995, delta_x JM = J - delta_x JP = J + delta_x T1 = CA1(IP, J) + CA1(I, JP) T2 = CA1(IM, J) + CA1(I, JM) S1 = T1 + T2 - 4 * CA1(I, J) CA(I, J) = CA1(I, J) + D * S1 end do end do In code 2, the culprit is conditionals do I = 1, N do J = 1, N If (IFLAG(I,J) .EQ. 0) then T1 = Value(I, J-1) T2 = Value(I-1, J) T3 = Value(I, J) T4 = Value(I+1, J) T5 = Value(I, J+1) Value(I,J) = 0.25 * (T1 + T2 + T5 + T4) Delta = ABS(T3 - Value(I,J)) If (Delta .GT. MaxDelta) MaxDelta = Delta endif enddo enddo I fixed both programs by inverting the loops by hand. Code 10 has three-dimensional arrays and triply nested loops. The structure of the most computationally intensive loops is too complex to invert automatically or by hand. The only practical solution is to transpose the arrays so that the dimension accessed by the innermost loop is in cache order. The arrays can be transposed at construction or prior to entering a computationally intensive section of code. The former requires all array references to be modified, while the latter is cost effective only if the cost of the transpose is amortized over many accesses. I used the second approach to optimize code 10. Code 5 has four-dimensional arrays and loops are nested four deep. For all of the reasons cited above the compiler is not able to restructure three key loops. Assume C arrays and let the four dimensions of the arrays be i, j, k, and l. In the original code, the index structure of the three loops is L1: for i L2: for i L3: for i for l for l for j for k for j for k for j for k for l So only L3 accesses array elements in cache order. L1 is a very complex loop—much too complex to invert. I brought the loop into cache alignment by transposing the second and fourth dimensions of the arrays. Since the code uses a macro to compute all array indexes, I effected the transpose at construction and changed the macro appropriately. The dimensions of the new arrays are now: i, l, k, and j. L3 is a simple loop and easily inverted. L2 has a loop-carried scalar dependence in k. By promoting the scalar name that carries the dependence to an array, I was able to invert the third and fourth subloops aligning the loop with cache. Code 5 is by far the most difficult of the four codes to optimize for array accesses; but the knowledge required to fix the problems is no more than that required for the other codes. I would judge this code at the limits of, but not beyond, the capabilities of appropriately trained computational scientists. Array Strides When a cache miss occurs, a line (64 bytes) rather than just one word is loaded into the cache. If data is accessed stride 1, than the cost of the miss is amortized over 8 words. Any stride other than one reduces the cost savings. Two of the ten codes studied suffered from non-unit strides. The codes represent two important classes of "strided" codes. Code 1 employs a multi-grid algorithm to reduce time to convergence. The grids are every tenth, fifth, second, and unit element. Since time to convergence is inversely proportional to the distance between elements, coarse grids converge quickly providing good starting values for finer grids. The better starting values further reduce the time to convergence. The downside is that grids of every nth element, n > 1, introduce non-unit strides into the computation. In the original code, much of the savings of the multi-grid algorithm were lost due to this problem. I eliminated the problem by compressing (copying) coarse grids into continuous memory, and rewriting the computation as a function of the compressed grid. On convergence, I copied the final values of the compressed grid back to the original grid. The savings gained from unit stride access of the compressed grid more than paid for the cost of copying. Using compressed grids, the loop from code 1 included in the previous section becomes do j = 1, GZ do i = 1, GZ T1 = CA(i+0, j-1) + CA(i-1, j+0) T4 = CA1(i+1, j+0) + CA1(i+0, j+1) S1 = T1 + T4 - 4 * CA1(i+0, j+0) CA(i+0, j+0) = CA1(i+0, j+0) + DD * S1 enddo enddo where CA and CA1 are compressed arrays of size GZ. Code 7 traverses a list of objects selecting objects for later processing. The labels of the selected objects are stored in an array. The selection step has unit stride, but the processing steps have irregular stride. A fix is to save the parameters of the selected objects in temporary arrays as they are selected, and pass the temporary arrays to the processing functions. The fix is practical if the same parameters are used in selection as in processing, or if processing comprises a series of distinct steps which use overlapping subsets of the parameters. Both conditions are true for code 7, so I achieved significant improvement by copying parameters to temporary arrays during selection. Data reuse In the previous sections, we optimized for spatial locality. It is also important to optimize for temporal locality. Once read, a datum should be used as much as possible before it is forced from cache. Loop fusion and loop unrolling are two techniques that increase temporal locality. Unfortunately, both techniques increase register pressure—as loop bodies become larger, the number of registers required to hold temporary values grows. Once register spilling occurs, any gains evaporate quickly. For multiprocessors with small register sets or small caches, the sweet spot can be very small. In the ten codes presented here, I found no opportunities for loop fusion and only two opportunities for loop unrolling (codes 1 and 3). In code 1, unrolling the outer and inner loop one iteration increases the number of result values computed by the loop body from 1 to 4, do J = 1, GZ-2, 2 do I = 1, GZ-2, 2 T1 = CA1(i+0, j-1) + CA1(i-1, j+0) T2 = CA1(i+1, j-1) + CA1(i+0, j+0) T3 = CA1(i+0, j+0) + CA1(i-1, j+1) T4 = CA1(i+1, j+0) + CA1(i+0, j+1) T5 = CA1(i+2, j+0) + CA1(i+1, j+1) T6 = CA1(i+1, j+1) + CA1(i+0, j+2) T7 = CA1(i+2, j+1) + CA1(i+1, j+2) S1 = T1 + T4 - 4 * CA1(i+0, j+0) S2 = T2 + T5 - 4 * CA1(i+1, j+0) S3 = T3 + T6 - 4 * CA1(i+0, j+1) S4 = T4 + T7 - 4 * CA1(i+1, j+1) CA(i+0, j+0) = CA1(i+0, j+0) + DD * S1 CA(i+1, j+0) = CA1(i+1, j+0) + DD * S2 CA(i+0, j+1) = CA1(i+0, j+1) + DD * S3 CA(i+1, j+1) = CA1(i+1, j+1) + DD * S4 enddo enddo The loop body executes 12 reads, whereas as the rolled loop shown in the previous section executes 20 reads to compute the same four values. In code 3, two loops are unrolled 8 times and one loop is unrolled 4 times. Here is the before for (k = 0; k < NK[u]; k++) { sum = 0.0; for (y = 0; y < NY; y++) { sum += W[y][u][k] * delta[y]; } backprop[i++]=sum; } and after code for (k = 0; k < KK - 8; k+=8) { sum0 = 0.0; sum1 = 0.0; sum2 = 0.0; sum3 = 0.0; sum4 = 0.0; sum5 = 0.0; sum6 = 0.0; sum7 = 0.0; for (y = 0; y < NY; y++) { sum0 += W[y][0][k+0] * delta[y]; sum1 += W[y][0][k+1] * delta[y]; sum2 += W[y][0][k+2] * delta[y]; sum3 += W[y][0][k+3] * delta[y]; sum4 += W[y][0][k+4] * delta[y]; sum5 += W[y][0][k+5] * delta[y]; sum6 += W[y][0][k+6] * delta[y]; sum7 += W[y][0][k+7] * delta[y]; } backprop[k+0] = sum0; backprop[k+1] = sum1; backprop[k+2] = sum2; backprop[k+3] = sum3; backprop[k+4] = sum4; backprop[k+5] = sum5; backprop[k+6] = sum6; backprop[k+7] = sum7; } for one of the loops unrolled 8 times. Optimizing for temporal locality is the most difficult optimization considered in this paper. The concepts are not difficult, but the sweet spot is small. Identifying where the program can benefit from loop unrolling or loop fusion is not trivial. Moreover, it takes some effort to get it right. Still, educating scientific programmers about temporal locality and teaching them how to optimize for it will pay dividends. Reducing instruction count Execution time is a function of instruction count. Reduce the count and you usually reduce the time. The best solution is to use a more efficient algorithm; that is, an algorithm whose order of complexity is smaller, that converges quicker, or is more accurate. Optimizing source code without changing the algorithm yields smaller, but still significant, gains. This paper considers only the latter because the intent is to study how much better codes can run if written by programmers schooled in basic code optimization techniques. The ten codes studied benefited from three types of "instruction reducing" optimizations. The two most prevalent were hoisting invariant memory and data operations out of inner loops. The third was eliminating unnecessary data copying. The nature of these inefficiencies is language dependent. Memory operations The semantics of C make it difficult for the compiler to determine all the invariant memory operations in a loop. The problem is particularly acute for loops in functions since the compiler may not know the values of the function's parameters at every call site when compiling the function. Most compilers support pragmas to help resolve ambiguities; however, these pragmas are not comprehensive and there is no standard syntax. To guarantee that invariant memory operations are not executed repetitively, the user has little choice but to hoist the operations by hand. The problem is not as severe in Fortran programs because in the absence of equivalence statements, it is a violation of the language's semantics for two names to share memory. Codes 3 and 5 are C programs. In both cases, the compiler did not hoist all invariant memory operations from inner loops. Consider the following loop from code 3 for (y = 0; y < NY; y++) { i = 0; for (u = 0; u < NU; u++) { for (k = 0; k < NK[u]; k++) { dW[y][u][k] += delta[y] * I1[i++]; } } } Since dW[y][u] can point to the same memory space as delta for one or more values of y and u, assignment to dW[y][u][k] may change the value of delta[y]. In reality, dW and delta do not overlap in memory, so I rewrote the loop as for (y = 0; y < NY; y++) { i = 0; Dy = delta[y]; for (u = 0; u < NU; u++) { for (k = 0; k < NK[u]; k++) { dW[y][u][k] += Dy * I1[i++]; } } } Failure to hoist invariant memory operations may be due to complex address calculations. If the compiler can not determine that the address calculation is invariant, then it can hoist neither the calculation nor the associated memory operations. As noted above, code 5 uses a macro to address four-dimensional arrays #define MAT4D(a,q,i,j,k) (double *)((a)->data + (q)*(a)->strides[0] + (i)*(a)->strides[3] + (j)*(a)->strides[2] + (k)*(a)->strides[1]) The macro is too complex for the compiler to understand and so, it does not identify any subexpressions as loop invariant. The simplest way to eliminate the address calculation from the innermost loop (over i) is to define a0 = MAT4D(a,q,0,j,k) before the loop and then replace all instances of *MAT4D(a,q,i,j,k) in the loop with a0[i] A similar problem appears in code 6, a Fortran program. The key loop in this program is do n1 = 1, nh nx1 = (n1 - 1) / nz + 1 nz1 = n1 - nz * (nx1 - 1) do n2 = 1, nh nx2 = (n2 - 1) / nz + 1 nz2 = n2 - nz * (nx2 - 1) ndx = nx2 - nx1 ndy = nz2 - nz1 gxx = grn(1,ndx,ndy) gyy = grn(2,ndx,ndy) gxy = grn(3,ndx,ndy) balance(n1,1) = balance(n1,1) + (force(n2,1) * gxx + force(n2,2) * gxy) * h1 balance(n1,2) = balance(n1,2) + (force(n2,1) * gxy + force(n2,2) * gyy)*h1 end do end do The programmer has written this loop well—there are no loop invariant operations with respect to n1 and n2. However, the loop resides within an iterative loop over time and the index calculations are independent with respect to time. Trading space for time, I precomputed the index values prior to the entering the time loop and stored the values in two arrays. I then replaced the index calculations with reads of the arrays. Data operations Ways to reduce data operations can appear in many forms. Implementing a more efficient algorithm produces the biggest gains. The closest I came to an algorithm change was in code 4. This code computes the inner product of K-vectors A(i) and B(j), 0 = i < N, 0 = j < M, for most values of i and j. Since the program computes most of the NM possible inner products, it is more efficient to compute all the inner products in one triply-nested loop rather than one at a time when needed. The savings accrue from reading A(i) once for all B(j) vectors and from loop unrolling. for (i = 0; i < N; i+=8) { for (j = 0; j < M; j++) { sum0 = 0.0; sum1 = 0.0; sum2 = 0.0; sum3 = 0.0; sum4 = 0.0; sum5 = 0.0; sum6 = 0.0; sum7 = 0.0; for (k = 0; k < K; k++) { sum0 += A[i+0][k] * B[j][k]; sum1 += A[i+1][k] * B[j][k]; sum2 += A[i+2][k] * B[j][k]; sum3 += A[i+3][k] * B[j][k]; sum4 += A[i+4][k] * B[j][k]; sum5 += A[i+5][k] * B[j][k]; sum6 += A[i+6][k] * B[j][k]; sum7 += A[i+7][k] * B[j][k]; } C[i+0][j] = sum0; C[i+1][j] = sum1; C[i+2][j] = sum2; C[i+3][j] = sum3; C[i+4][j] = sum4; C[i+5][j] = sum5; C[i+6][j] = sum6; C[i+7][j] = sum7; }} This change requires knowledge of a typical run; i.e., that most inner products are computed. The reasons for the change, however, derive from basic optimization concepts. It is the type of change easily made at development time by a knowledgeable programmer. In code 5, we have the data version of the index optimization in code 6. Here a very expensive computation is a function of the loop indices and so cannot be hoisted out of the loop; however, the computation is invariant with respect to an outer iterative loop over time. We can compute its value for each iteration of the computation loop prior to entering the time loop and save the values in an array. The increase in memory required to store the values is small in comparison to the large savings in time. The main loop in Code 8 is doubly nested. The inner loop includes a series of guarded computations; some are a function of the inner loop index but not the outer loop index while others are a function of the outer loop index but not the inner loop index for (j = 0; j < N; j++) { for (i = 0; i < M; i++) { r = i * hrmax; R = A[j]; temp = (PRM[3] == 0.0) ? 1.0 : pow(r, PRM[3]); high = temp * kcoeff * B[j] * PRM[2] * PRM[4]; low = high * PRM[6] * PRM[6] / (1.0 + pow(PRM[4] * PRM[6], 2.0)); kap = (R > PRM[6]) ? high * R * R / (1.0 + pow(PRM[4]*r, 2.0) : low * pow(R/PRM[6], PRM[5]); < rest of loop omitted > }} Note that the value of temp is invariant to j. Thus, we can hoist the computation for temp out of the loop and save its values in an array. for (i = 0; i < M; i++) { r = i * hrmax; TEMP[i] = pow(r, PRM[3]); } [N.B. – the case for PRM[3] = 0 is omitted and will be reintroduced later.] We now hoist out of the inner loop the computations invariant to i. Since the conditional guarding the value of kap is invariant to i, it behooves us to hoist the computation out of the inner loop, thereby executing the guard once rather than M times. The final version of the code is for (j = 0; j < N; j++) { R = rig[j] / 1000.; tmp1 = kcoeff * par[2] * beta[j] * par[4]; tmp2 = 1.0 + (par[4] * par[4] * par[6] * par[6]); tmp3 = 1.0 + (par[4] * par[4] * R * R); tmp4 = par[6] * par[6] / tmp2; tmp5 = R * R / tmp3; tmp6 = pow(R / par[6], par[5]); if ((par[3] == 0.0) && (R > par[6])) { for (i = 1; i <= imax1; i++) KAP[i] = tmp1 * tmp5; } else if ((par[3] == 0.0) && (R <= par[6])) { for (i = 1; i <= imax1; i++) KAP[i] = tmp1 * tmp4 * tmp6; } else if ((par[3] != 0.0) && (R > par[6])) { for (i = 1; i <= imax1; i++) KAP[i] = tmp1 * TEMP[i] * tmp5; } else if ((par[3] != 0.0) && (R <= par[6])) { for (i = 1; i <= imax1; i++) KAP[i] = tmp1 * TEMP[i] * tmp4 * tmp6; } for (i = 0; i < M; i++) { kap = KAP[i]; r = i * hrmax; < rest of loop omitted > } } Maybe not the prettiest piece of code, but certainly much more efficient than the original loop, Copy operations Several programs unnecessarily copy data from one data structure to another. This problem occurs in both Fortran and C programs, although it manifests itself differently in the two languages. Code 1 declares two arrays—one for old values and one for new values. At the end of each iteration, the array of new values is copied to the array of old values to reset the data structures for the next iteration. This problem occurs in Fortran programs not included in this study and in both Fortran 77 and Fortran 90 code. Introducing pointers to the arrays and swapping pointer values is an obvious way to eliminate the copying; but pointers is not a feature that many Fortran programmers know well or are comfortable using. An easy solution not involving pointers is to extend the dimension of the value array by 1 and use the last dimension to differentiate between arrays at different times. For example, if the data space is N x N, declare the array (N, N, 2). Then store the problem’s initial values in (_, _, 2) and define the scalar names new = 2 and old = 1. At the start of each iteration, swap old and new to reset the arrays. The old–new copy problem did not appear in any C program. In programs that had new and old values, the code swapped pointers to reset data structures. Where unnecessary coping did occur is in structure assignment and parameter passing. Structures in C are handled much like scalars. Assignment causes the data space of the right-hand name to be copied to the data space of the left-hand name. Similarly, when a structure is passed to a function, the data space of the actual parameter is copied to the data space of the formal parameter. If the structure is large and the assignment or function call is in an inner loop, then copying costs can grow quite large. While none of the ten programs considered here manifested this problem, it did occur in programs not included in the study. A simple fix is always to refer to structures via pointers. Optimizing loop structures Since scientific programs spend almost all their time in loops, efficient loops are the key to good performance. Conditionals, function calls, little instruction level parallelism, and large numbers of temporary values make it difficult for the compiler to generate tightly packed, highly efficient code. Conditionals and function calls introduce jumps that disrupt code flow. Users should eliminate or isolate conditionls to their own loops as much as possible. Often logical expressions can be substituted for if-then-else statements. For example, code 2 includes the following snippet MaxDelta = 0.0 do J = 1, N do I = 1, M < code omitted > Delta = abs(OldValue ? NewValue) if (Delta > MaxDelta) MaxDelta = Delta enddo enddo if (MaxDelta .gt. 0.001) goto 200 Since the only use of MaxDelta is to control the jump to 200 and all that matters is whether or not it is greater than 0.001, I made MaxDelta a boolean and rewrote the snippet as MaxDelta = .false. do J = 1, N do I = 1, M < code omitted > Delta = abs(OldValue ? NewValue) MaxDelta = MaxDelta .or. (Delta .gt. 0.001) enddo enddo if (MaxDelta) goto 200 thereby, eliminating the conditional expression from the inner loop. A microprocessor can execute many instructions per instruction cycle. Typically, it can execute one or more memory, floating point, integer, and jump operations. To be executed simultaneously, the operations must be independent. Thick loops tend to have more instruction level parallelism than thin loops. Moreover, they reduce memory traffice by maximizing data reuse. Loop unrolling and loop fusion are two techniques to increase the size of loop bodies. Several of the codes studied benefitted from loop unrolling, but none benefitted from loop fusion. This observation is not too surpising since it is the general tendency of programmers to write thick loops. As loops become thicker, the number of temporary values grows, increasing register pressure. If registers spill, then memory traffic increases and code flow is disrupted. A thick loop with many temporary values may execute slower than an equivalent series of thin loops. The biggest gain will be achieved if the thick loop can be split into a series of independent loops eliminating the need to write and read temporary arrays. I found such an occasion in code 10 where I split the loop do i = 1, n do j = 1, m A24(j,i)= S24(j,i) * T24(j,i) + S25(j,i) * U25(j,i) B24(j,i)= S24(j,i) * T25(j,i) + S25(j,i) * U24(j,i) A25(j,i)= S24(j,i) * C24(j,i) + S25(j,i) * V24(j,i) B25(j,i)= S24(j,i) * U25(j,i) + S25(j,i) * V25(j,i) C24(j,i)= S26(j,i) * T26(j,i) + S27(j,i) * U26(j,i) D24(j,i)= S26(j,i) * T27(j,i) + S27(j,i) * V26(j,i) C25(j,i)= S27(j,i) * S28(j,i) + S26(j,i) * U28(j,i) D25(j,i)= S27(j,i) * T28(j,i) + S26(j,i) * V28(j,i) end do end do into two disjoint loops do i = 1, n do j = 1, m A24(j,i)= S24(j,i) * T24(j,i) + S25(j,i) * U25(j,i) B24(j,i)= S24(j,i) * T25(j,i) + S25(j,i) * U24(j,i) A25(j,i)= S24(j,i) * C24(j,i) + S25(j,i) * V24(j,i) B25(j,i)= S24(j,i) * U25(j,i) + S25(j,i) * V25(j,i) end do end do do i = 1, n do j = 1, m C24(j,i)= S26(j,i) * T26(j,i) + S27(j,i) * U26(j,i) D24(j,i)= S26(j,i) * T27(j,i) + S27(j,i) * V26(j,i) C25(j,i)= S27(j,i) * S28(j,i) + S26(j,i) * U28(j,i) D25(j,i)= S27(j,i) * T28(j,i) + S26(j,i) * V28(j,i) end do end do Conclusions Over the course of the last year, I have had the opportunity to work with over two dozen academic scientific programmers at leading research universities. Their research interests span a broad range of scientific fields. Except for two programs that relied almost exclusively on library routines (matrix multiply and fast Fourier transform), I was able to improve significantly the single processor performance of all codes. Improvements range from 2x to 15.5x with a simple average of 4.75x. Changes to the source code were at a very high level. I did not use sophisticated techniques or programming tools to discover inefficiencies or effect the changes. Only one code was parallel despite the availability of parallel systems to all developers. Clearly, we have a problem—personal scientific research codes are highly inefficient and not running parallel. The developers are unaware of simple optimization techniques to make programs run faster. They lack education in the art of code optimization and parallel programming. I do not believe we can fix the problem by publishing additional books or training manuals. To date, the developers in questions have not studied the books or manual available, and are unlikely to do so in the future. Short courses are a possible solution, but I believe they are too concentrated to be much use. The general concepts can be taught in a three or four day course, but that is not enough time for students to practice what they learn and acquire the experience to apply and extend the concepts to their codes. Practice is the key to becoming proficient at optimization. I recommend that graduate students be required to take a semester length course in optimization and parallel programming. We would never give someone access to state-of-the-art scientific equipment costing hundreds of thousands of dollars without first requiring them to demonstrate that they know how to use the equipment. Yet the criterion for time on state-of-the-art supercomputers is at most an interesting project. Requestors are never asked to demonstrate that they know how to use the system, or can use the system effectively. A semester course would teach them the required skills. Government agencies that fund academic scientific research pay for most of the computer systems supporting scientific research as well as the development of most personal scientific codes. These agencies should require graduate schools to offer a course in optimization and parallel programming as a requirement for funding. About the Author John Feo received his Ph.D. in Computer Science from The University of Texas at Austin in 1986. After graduate school, Dr. Feo worked at Lawrence Livermore National Laboratory where he was the Group Leader of the Computer Research Group and principal investigator of the Sisal Language Project. In 1997, Dr. Feo joined Tera Computer Company where he was project manager for the MTA, and oversaw the programming and evaluation of the MTA at the San Diego Supercomputer Center. In 2000, Dr. Feo joined Sun Microsystems as an HPC application specialist. He works with university research groups to optimize and parallelize scientific codes. Dr. Feo has published over two dozen research articles in the areas of parallel parallel programming, parallel programming languages, and application performance.

    Read the article

  • C#/.NET Little Wonders: Fun With Enum Methods

    - by James Michael Hare
    Once again lets dive into the Little Wonders of .NET, those small things in the .NET languages and BCL classes that make development easier by increasing readability, maintainability, and/or performance. So probably every one of us has used an enumerated type at one time or another in a C# program.  The enumerated types we create are a great way to represent that a value can be one of a set of discrete values (or a combination of those values in the case of bit flags). But the power of enum types go far beyond simple assignment and comparison, there are many methods in the Enum class (that all enum types “inherit” from) that can give you even more power when dealing with them. IsDefined() – check if a given value exists in the enum Are you reading a value for an enum from a data source, but are unsure if it is actually a valid value or not?  Casting won’t tell you this, and Parse() isn’t guaranteed to balk either if you give it an int or a combination of flags.  So what can we do? Let’s assume we have a small enum like this for result codes we want to return back from our business logic layer: 1: public enum ResultCode 2: { 3: Success, 4: Warning, 5: Error 6: } In this enum, Success will be zero (unless given another value explicitly), Warning will be one, and Error will be two. So what happens if we have code like this where perhaps we’re getting the result code from another data source (could be database, could be web service, etc)? 1: public ResultCode PerformAction() 2: { 3: // set up and call some method that returns an int. 4: int result = ResultCodeFromDataSource(); 5:  6: // this will suceed even if result is < 0 or > 2. 7: return (ResultCode) result; 8: } So what happens if result is –1 or 4?  Well, the cast does not fail, so what we end up with would be an instance of a ResultCode that would have a value that’s outside of the bounds of the enum constants we defined. This means if you had a block of code like: 1: switch (result) 2: { 3: case ResultType.Success: 4: // do success stuff 5: break; 6:  7: case ResultType.Warning: 8: // do warning stuff 9: break; 10:  11: case ResultType.Error: 12: // do error stuff 13: break; 14: } That you would hit none of these blocks (which is a good argument for always having a default in a switch by the way). So what can you do?  Well, there is a handy static method called IsDefined() on the Enum class which will tell you if an enum value is defined.  1: public ResultCode PerformAction() 2: { 3: int result = ResultCodeFromDataSource(); 4:  5: if (!Enum.IsDefined(typeof(ResultCode), result)) 6: { 7: throw new InvalidOperationException("Enum out of range."); 8: } 9:  10: return (ResultCode) result; 11: } In fact, this is often recommended after you Parse() or cast a value to an enum as there are ways for values to get past these methods that may not be defined. If you don’t like the syntax of passing in the type of the enum, you could clean it up a bit by creating an extension method instead that would allow you to call IsDefined() off any isntance of the enum: 1: public static class EnumExtensions 2: { 3: // helper method that tells you if an enum value is defined for it's enumeration 4: public static bool IsDefined(this Enum value) 5: { 6: return Enum.IsDefined(value.GetType(), value); 7: } 8: }   HasFlag() – an easier way to see if a bit (or bits) are set Most of us who came from the land of C programming have had to deal extensively with bit flags many times in our lives.  As such, using bit flags may be almost second nature (for a quick refresher on bit flags in enum types see one of my old posts here). However, in higher-level languages like C#, the need to manipulate individual bit flags is somewhat diminished, and the code to check for bit flag enum values may be obvious to an advanced developer but cryptic to a novice developer. For example, let’s say you have an enum for a messaging platform that contains bit flags: 1: // usually, we pluralize flags enum type names 2: [Flags] 3: public enum MessagingOptions 4: { 5: None = 0, 6: Buffered = 0x01, 7: Persistent = 0x02, 8: Durable = 0x04, 9: Broadcast = 0x08 10: } We can combine these bit flags using the bitwise OR operator (the ‘|’ pipe character): 1: // combine bit flags using 2: var myMessenger = new Messenger(MessagingOptions.Buffered | MessagingOptions.Broadcast); Now, if we wanted to check the flags, we’d have to test then using the bit-wise AND operator (the ‘&’ character): 1: if ((options & MessagingOptions.Buffered) == MessagingOptions.Buffered) 2: { 3: // do code to set up buffering... 4: // ... 5: } While the ‘|’ for combining flags is easy enough to read for advanced developers, the ‘&’ test tends to be easy for novice developers to get wrong.  First of all you have to AND the flag combination with the value, and then typically you should test against the flag combination itself (and not just for a non-zero)!  This is because the flag combination you are testing with may combine multiple bits, in which case if only one bit is set, the result will be non-zero but not necessarily all desired bits! Thanks goodness in .NET 4.0 they gave us the HasFlag() method.  This method can be called from an enum instance to test to see if a flag is set, and best of all you can avoid writing the bit wise logic yourself.  Not to mention it will be more readable to a novice developer as well: 1: if (options.HasFlag(MessagingOptions.Buffered)) 2: { 3: // do code to set up buffering... 4: // ... 5: } It is much more concise and unambiguous, thus increasing your maintainability and readability. It would be nice to have a corresponding SetFlag() method, but unfortunately generic types don’t allow you to specialize on Enum, which makes it a bit more difficult.  It can be done but you have to do some conversions to numeric and then back to the enum which makes it less of a payoff than having the HasFlag() method.  But if you want to create it for symmetry, it would look something like this: 1: public static T SetFlag<T>(this Enum value, T flags) 2: { 3: if (!value.GetType().IsEquivalentTo(typeof(T))) 4: { 5: throw new ArgumentException("Enum value and flags types don't match."); 6: } 7:  8: // yes this is ugly, but unfortunately we need to use an intermediate boxing cast 9: return (T)Enum.ToObject(typeof (T), Convert.ToUInt64(value) | Convert.ToUInt64(flags)); 10: } Note that since the enum types are value types, we need to assign the result to something (much like string.Trim()).  Also, you could chain several SetFlag() operations together or create one that takes a variable arg list if desired. Parse() and ToString() – transitioning from string to enum and back Sometimes, you may want to be able to parse an enum from a string or convert it to a string - Enum has methods built in to let you do this.  Now, many may already know this, but may not appreciate how much power are in these two methods. For example, if you want to parse a string as an enum, it’s easy and works just like you’d expect from the numeric types: 1: string optionsString = "Persistent"; 2:  3: // can use Enum.Parse, which throws if finds something it doesn't like... 4: var result = (MessagingOptions)Enum.Parse(typeof (MessagingOptions), optionsString); 5:  6: if (result == MessagingOptions.Persistent) 7: { 8: Console.WriteLine("It worked!"); 9: } Note that Enum.Parse() will throw if it finds a value it doesn’t like.  But the values it likes are fairly flexible!  You can pass in a single value, or a comma separated list of values for flags and it will parse them all and set all bits: 1: // for string values, can have one, or comma separated. 2: string optionsString = "Persistent, Buffered"; 3:  4: var result = (MessagingOptions)Enum.Parse(typeof (MessagingOptions), optionsString); 5:  6: if (result.HasFlag(MessagingOptions.Persistent) && result.HasFlag(MessagingOptions.Buffered)) 7: { 8: Console.WriteLine("It worked!"); 9: } Or you can parse in a string containing a number that represents a single value or combination of values to set: 1: // 3 is the combination of Buffered (0x01) and Persistent (0x02) 2: var optionsString = "3"; 3:  4: var result = (MessagingOptions) Enum.Parse(typeof (MessagingOptions), optionsString); 5:  6: if (result.HasFlag(MessagingOptions.Persistent) && result.HasFlag(MessagingOptions.Buffered)) 7: { 8: Console.WriteLine("It worked again!"); 9: } And, if you really aren’t sure if the parse will work, and don’t want to handle an exception, you can use TryParse() instead: 1: string optionsString = "Persistent, Buffered"; 2: MessagingOptions result; 3:  4: // try parse returns true if successful, and takes an out parm for the result 5: if (Enum.TryParse(optionsString, out result)) 6: { 7: if (result.HasFlag(MessagingOptions.Persistent) && result.HasFlag(MessagingOptions.Buffered)) 8: { 9: Console.WriteLine("It worked!"); 10: } 11: } So we covered parsing a string to an enum, what about reversing that and converting an enum to a string?  The ToString() method is the obvious and most basic choice for most of us, but did you know you can pass a format string for enum types that dictate how they are written as a string?: 1: MessagingOptions value = MessagingOptions.Buffered | MessagingOptions.Persistent; 2:  3: // general format, which is the default, 4: Console.WriteLine("Default : " + value); 5: Console.WriteLine("G (default): " + value.ToString("G")); 6:  7: // Flags format, even if type does not have Flags attribute. 8: Console.WriteLine("F (flags) : " + value.ToString("F")); 9:  10: // integer format, value as number. 11: Console.WriteLine("D (num) : " + value.ToString("D")); 12:  13: // hex format, value as hex 14: Console.WriteLine("X (hex) : " + value.ToString("X")); Which displays: 1: Default : Buffered, Persistent 2: G (default): Buffered, Persistent 3: F (flags) : Buffered, Persistent 4: D (num) : 3 5: X (hex) : 00000003 Now, you may not really see a difference here between G and F because I used a [Flags] enum, the difference is that the “F” option treats the enum as if it were flags even if the [Flags] attribute is not present.  Let’s take a non-flags enum like the ResultCode used earlier: 1: // yes, we can do this even if it is not [Flags] enum. 2: ResultCode value = ResultCode.Warning | ResultCode.Error; And if we run that through the same formats again we get: 1: Default : 3 2: G (default): 3 3: F (flags) : Warning, Error 4: D (num) : 3 5: X (hex) : 00000003 Notice that since we had multiple values combined, but it was not a [Flags] marked enum, the G and default format gave us a number instead of a value name.  This is because the value was not a valid single-value constant of the enum.  However, using the F flags format string, it broke out the value into its component flags even though it wasn’t marked [Flags]. So, if you want to get an enum to display appropriately for whether or not it has the [Flags] attribute, use G which is the default.  If you always want it to attempt to break down the flags, use F.  For numeric output, obviously D or  X are the best choice depending on whether you want decimal or hex. Summary Hopefully, you learned a couple of new tricks with using the Enum class today!  I’ll add more little wonders as I think of them and thanks for all the invaluable input!   Technorati Tags: C#,.NET,Little Wonders,Enum,BlackRabbitCoder

    Read the article

  • Why 32-bit color EGL configurations fail with EGL_BAD_MATCH on Moto Droid?

    - by Gilead
    I'm trying to figure out why certain EGL configurations cause eglMakeCurrent() call to return EGL_BAD_MATCH on Motorola Droid running Android 2.1u1. This is a full list of hardware-accelerated EGL configurations (those with EGL_CONFIG_CAVEAT == EGL_NONE) as there's a few others with EGL_CONFIG_CAVEAT == EGL_SLOW_CONFIG but those are backed by PixelFlinger 1.2 meaning they're using software renderer. ID: 0 RGB: 8, 8, 8 Alpha: 8 Depth: 24 Stencil: 8 // BAD MATCH ID: 1 RGB: 8, 8, 8 Alpha: 8 Depth: 0 Stencil: 0 // BAD MATCH ID: 2 RGB: 8, 8, 8 Alpha: 8 Depth: 24 Stencil: 8 // BAD MATCH ID: 3 RGB: 8, 8, 8 Alpha: 8 Depth: 24 Stencil: 8 // BAD MATCH ID: 4 RGB: 8, 8, 8 Alpha: 8 Depth: 0 Stencil: 0 // BAD MATCH ID: 5 RGB: 8, 8, 8 Alpha: 8 Depth: 24 Stencil: 8 // BAD MATCH ID: 6 RGB: 5, 6, 5 Alpha: 0 Depth: 24 Stencil: 8 ID: 7 RGB: 5, 6, 5 Alpha: 0 Depth: 0 Stencil: 0 ID: 8 RGB: 5, 6, 5 Alpha: 0 Depth: 24 Stencil: 8 Clearly, all configurations with 32-bit color depth fail and all 16-bit ones are OK but: 1. Why? 2. WHY?! :) 3. How do I tell which ones would fail before actually trying to use them? The code below is as simple as it can get. I put if (v[0] == 6) there to check different configs, normally they're chosen by half-clever config matcher :) private void createSurface(SurfaceHolder holder) { egl = (EGL10)EGLContext.getEGL(); eglDisplay = egl.eglGetDisplay(EGL10.EGL_DEFAULT_DISPLAY); egl.eglInitialize(eglDisplay, null); int[] numConfigs = new int[1]; egl.eglChooseConfig(eglDisplay, new int[] { EGL10.EGL_NONE }, null, 0, numConfigs); EGLConfig[] configs = new EGLConfig[numConfigs[0]]; egl.eglChooseConfig(eglDisplay, new int[] { EGL10.EGL_NONE }, configs, numConfigs[0], numConfigs); int[] v = new int[1]; for (EGLConfig c : configs) { egl.eglGetConfigAttrib(eglDisplay, c, EGL10.EGL_CONFIG_ID, v); if (v[0] == 6) { eglConfig = c; } } eglContext = egl.eglCreateContext(eglDisplay, eglConfig, EGL10.EGL_NO_CONTEXT, null); if (eglContext == null || eglContext == EGL10.EGL_NO_CONTEXT) { throw new RuntimeException("Unable to create EGL context"); } eglSurface = egl.eglCreateWindowSurface(eglDisplay, eglConfig, holder, null); if (eglSurface == null || eglSurface == EGL10.EGL_NO_SURFACE) { throw new RuntimeException("Unable to create EGL surface"); } if (!egl.eglMakeCurrent(eglDisplay, eglSurface, eglSurface, eglContext)) { throw new RuntimeException("Unable to make EGL current"); } gl = (GL10)eglContext.getGL(); }

    Read the article

  • Any downsides to UPX-ing my 32-bit Python 2.6.4 development environment EXE/PYD/DLL files?

    - by Malcolm
    Are there any downsides to UPX-ing my 32-bit Python 2.6.4 development environment EXE/PYD/DLL files? The reason I'm asking is that I frequently use a custom PY2EXE script that UPX's copies of these files on every build. Yes, I could get fancy and try to cache UPXed files, but I think a simpler, safer, and higher performance solution would be for me to just UPX my Python 2.6.4 directory once and be done with it. Thoughts? Malcolm

    Read the article

  • Microsoft Windows 64-bit application development best practises installation folder.

    - by abmv
    My problem is that a vendor is providing me with a 64bit application (packed in a 64bit installer) but it goes and installs to the x86 (Program Files) Folder and he keeps telling me its OK but I want it to install in the Program Files directory; as the 32 bit version does that and scripts for the app are developed based on this assumption. Can someone direct me to the Microsoft recommended best practices for 64bit applications(links). Thanks in advance.

    Read the article

  • What alternatives to __attribute__ exist on 64-bit kernels?

    - by Saifi Khan
    Hi: Is there any alternative to non-ISO gcc specific extension __attribute__ on 64-bit kernels ? Three types that i've noticed are: function attributes, type attributes and variable attributes. eg. i'd like to avoid using __attribute__((__packed__)) for structures passed over the network, even though some gcc based code do use it. Any suggestions or pointers on how to entirely avoid __attribute__ usage in C systems/kernel code ? thanks Saifi.

    Read the article

< Previous Page | 82 83 84 85 86 87 88 89 90 91 92 93  | Next Page >