Search Results

Search found 12 results on 1 pages for 'kap'.

Page 1/1 | 1 

  • 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

  • Casting To The Correct Subclass

    - by kap
    Hi Guys I hava a supeclass called Car with 3 subclasses. class Ford extends Car{ } class Chevrolet extends Car{ } class Audi extends Car{ } Now i have a function called printMessge(Car car) which will print a message of a particular car type. In the implementation i use if statements to test the instance of the classes like this. public int printMessge(Car car){ if((Ford)car instanceof Ford){ // print ford }else if((Chevrolet)car instanceof Chevrolet){ // print chevrolet }else if((Audi)car instanceof Audi){ // print Audi } } for instance if i call it for the first time with Ford printMessge(new Ford()), it prints the ford message but when i call it with printMessge(new Chevrolet()), i get EXCEPTION from the first if statement that Chevrolet cannot be cast to Ford. What am i doing wrong and what is the best way. thanks

    Read the article

  • Javax Swing Timer Help

    - by kap
    Hello Guys, I am having some problems concerning starting javax.swing.Timer after a mouse click. I want to start the timer to perform some animation after the user clicks on a button but it is not working. Here are the code snippets: public class ShowMe extends JPanel{ private javax.swing.Timer timer; public ShowMe(){ timer = new javax.swing.Timer(20, new MoveListener()); } // getters and setters here private class MoveListener implements ActionListener { public void actionPerformed(ActionEvent e) { // some code here to perform the animation } } } This is the class which contains a button so that when the user clicks on the button the timer starts to begin the animation public class Test{ // button declarations go here and registering listeners also here public void actionPerformed(ActionEvent e) { if(e.getSource() == this.btnConnect){ ShowMe vis = new ShowMe(); vis.getTimer().start(); } } } I want to start the timer to begin the animation but it is not working. Need help how to make a timer start after button click. Thanks.

    Read the article

  • Html5 - Callback when media is ready on iPad wont work

    - by Kap
    I'm trying to add a callback to a HTML5 audio element on an iPad. I added an eventlistener to the element, the myOtherThing() starts but there is no sound. If I pause and the play the sound again the audio starts. This works in chrome. Does anyone have an idea how I can do this? myAudioElement.src = "path_to_file"; addEventListener("canplay", function(){ myAudioElement.play(); myOtherThing.start(); }); SOLVED Just wanted to share my solution here, just in case someone else needs it. As far as I understand the iPad does not trigger any events without user interactions. So to be able to use "canply", "playing" and all the other events you need to use the built in media controller. Once you press play in that controller, the events gets triggered. After that you can use your custom interface.

    Read the article

  • Displaying Current Pictures From A Camera

    - by Kap
    Hi Guys, I would like to develop an application that will receive pictures from a camera and maybe afterwards save it in a database. This is what i want to do: When the picture is taken it is send to the program immediately (or the program must read the current picture taken) then display it. I will take pictures of many people or things so anytime a picture is taken i want to see the current picture displayed in the program. I have googled if i can see an example application so that i know that it possible so that i can do mine from scratch. But couldn't find any so i am not sure if it is possible to do it in java. So guys am asking for guidelines how i can do it in java. I just need the steps then i will program everything myself. Thanks.

    Read the article

  • JList strike through

    - by kap
    I have a list of data in a JList component in my GUI. I would like to know if there is a method that i can call on the list element(s) to strike through a particular element in the list. I would like to draw a line through the element to appear as if that element is canceled. I want a similar thing like the strike through functionality in Microsoft Word document whereby a line i drawn through the text. thanks for your help

    Read the article

  • python dictionary with constant value-type

    - by s.kap
    hi there, I bumped into a case where I need a big (=huge) python dictionary, which turned to be quite memory-consuming. However, since all of the values are of a single type (long) - as well as the keys, I figured I can use python (or numpy, doesn't really matter) array for the values ; and wrap the needed interface (in: x ; out: d[x]) with an object which actually uses these arrays for the keys and values storage. I can use a index-conversion object (input -- index, of 1..n, where n is the different-values counter), and return array[index]. I can elaborate on some techniques of how to implement such an indexing-methods with reasonable memory requirement, it works and even pretty good. However, I wonder if there is such a data-structure-object already exists (in python, or wrapped to python from C/++), in any package (I checked collections, and some Google searches). Any comment will be welcome, thanks.

    Read the article

  • Laptoppal a HOUG konferenci&aacute;ra

    - by Lajos Sárecz
    Mához 3 hétre kezdodik a HOUG konferencia. Március 28-án hétfon, a konferencia 0. napján délután Workshop-okkal indítjuk a konferenciát, amelyek közül több is lehetoséget ad arra, hogy a résztvevok saját laptopjukon kipróbálhassák az Oracle különbözo termékeit. Én egy Oracle Data Masking Hands-on Workshop-ot fogok tartani a deperszonalizáció, anonimizálás bemutatására, amely keretében egy Virtualbox image-et kap minden résztvevo. Szükség lesz kb. 20GB szabad területre, 3 GB memóriára. Valami oknál fogva a Data Masking Demo nem szereti az AMD processzorokat, így érdemes Intel alapú processzorral felszerelt laptoppal érkezni. Mivel egymás után több hands-on részvételre is lehetoség nyílik, ezért aki szeretné az image-eket megorízni, az készüljön megfelelo méretu háttértárral.

    Read the article

  • MAA a Database Machine-nel, maximális rendelkezésre állás

    - by Fekete Zoltán
    Néhány napja jelent meg egy, a maximális rendelkezésre állást boncolgató Oracle fehérpapír :): Oracle Data Guard: Disaster Recovery for Sun Oracle Database Machine. Ez a dokumentum az Exadata környezetben az Oracle Data Guard használatát elemzi. Az utolsó oldalakon néhány rendkívül hasznos linket is találunk. Mire is használható a Data Guard? - katasztrófa helyzet kezelése - adatbázis gördülo upgrade - egy megoldás az Exadata környezetre migrálásra - a standby adatbázis kihasználása A Sun Oracle Database Machine háromféle konfigurációban kapható: Full Rack, Half Rack és Quarter Rack, azaz teljes, fél és negyed szekrény kiépítésben. Felfelé upgrade-elheto és akár sok Full Rack összekapcsolva is egyetlen gépként tud muködni. A határ tehát a csillagos ég! :) Hiszen a nap a legfontosabb csillagunk. A Database Machine már önmagában is magas rendelkezésreállást biztosít, hiszen minden - a muködés szempontjából fontos - minden komponense legalább duplikált! Természetesen ez az adatokra is vonatkozik. A Database Machine ideális gyors környezet mind OLTP, mind DW futtatására, mind adatbázis konszolidációra. A tranzakciós (OLTP) rendszereknél régóta fontos követelmény, hogy az elsodleges site mögött legyen egy katasztrófa site, mely át tudja venni az adatbázis-kezelés feladatát, ha árvíz, tuz, vagy más szomorú katasztrófa történne az elsodleges site-on. Manapság már az adattárházak (DW) üzemeltetésében is fontos szerepet kap az MAA architektúra, azaz a Maximum Availability Architecture. Innen letöltheto a pdf: Oracle Data Guard: Disaster Recovery for Sun Oracle Database Machine.

    Read the article

  • Adatbázis szerver konszolidáció Oracle technológiákkal - eroforrás allokálás

    - by Lajos Sárecz
    Szerver konszolidációnál alapmegoldás a virtualizáció, pedig az Oracle Database rendelkezik olyan képességekkel, melyekkel a virtualizáció elonyeit élvezhetjük, ám teljesítményben felülmúljuk azt. Több adatbázis konszolidációját meg lehet oldani egy nagy szerveren, vagy egy több szerverbol álló klaszteren. Bármelyik megoldást is választjuk (ezek elonyeivel és hátrányaival most nem foglalkozok), az egyik legfontosabb megoldandó probléma, hogy biztonsággal el tudjuk oket szeparálni akár adatbiztonsági, akár eroforrás kezelési szempontból. A szoftveres és hardveres virtualizációk lehetové teszik, hogy a szerver eroforrásait több virtuális szerver között felosszuk, ezáltal elszeparálhatók a párhuzamosan futó adatbázis példányok. Ezek a megoldások általában költségesek, plusz adminisztrációt jelentenek és teljesítmény csökkenést okoznak. Az alábbiakban röviden összeszedem, hogy az Oracle Database milyen eroforrás szeparációs technológiákkal rendelkezik, melyek jól használhatók adatbázis konszolidáció esetén: Adatbázis szolgáltatások: Azt talán minden Oracle adatbázis-kezelovel foglalkozó szakérto tudja, hogy akliensek az adatbázist az adatbázis szolgáltatás nevével érik el. Alapértelmezetten minden adatbázis egyetlen szolgáltatással rendelkezik, mely automatikusan a 'global database name' paraméterrel megegyezo nevet kapja az adatbázis létrehozásakor. Ugyanakkor egy adatbázishoz több szolgáltatás név is rendelheto. A szolgáltatásokkal csoportosíthatók a különbözo feladatokat végrehajtó kliensek, és a szolgáltatásokhoz rendelhetjük hogy melyik kliens csoportnak mennyi rendszer eroforrást allokálunk. Klaszteres adatbázisok (RAC) esetén egy szolgáltatás több adatbázis példányhoz (szerverhez) kapcsolódhat, amivel valós terheléstol függo terhelés elosztás valósítható meg (itt már szerepet kap egyébként a Resource Manager is, lásd késobb). Az alkalmazás számára irrelevánssá válik, hogy az adott szolgáltatást mely szerver szolgálja ki. A szolgáltatásokhoz kapcsolódó eroforrások menet közben dinamikusan bovíthetok, de kezelik a kieso eroforrások hiányát is (failover). Database Resource Manager: Az Oracle Database Resource Manager az adatbázis szintjén kezeli az eroforrásokat, a CPU használatot szabályozza az adatbázis terhelés kontrolljával. A Resource Manager egy CPU-n adott pillanatban csak egyetlen Oracle processz futtatását engedélyezi, miközben a többit várakoztatja (ahogy az egy operációs rendszer ütemezojében is muködik). A Resource Manager csak akkor lép muködésbe, amikor a CPU terhelése eléri a 100%-ot. Ekkor a Resource Plan-nek megfeleloen korlátozhatja az egyes eroforrás csoportok számára elérheto eroforrás (CPU) mennyiségét. Instance Caging: A Resource Manager részeként az Oracle Database 11gR2-tol elérheto Instance Caging technológiával virtualizáció és operációs rendszer szintu eroforrás felosztás nélkül az adatbázis példány szintjén lehet szabályozni az allokált CPU számot. Erre akkor lehet szükség, ha egy szerveren több példány futtatására van szükség. A Resource Manager bekapcsolásával és a cpu_count paraméter beállításával lehet adatbázis példányonként aktiválni az Instance Caging funkcionalitást. A cpu_count egy dinamikus paraméter, célszeru arra az értékre állítani, ahány CPU-t az adott adatbázis példány maximálisan igényelhet. Lehetoség van túlméretezni a példányok számára rendelkezésre álló processzorok számát. Például egy 4 CPUs- szerver esetében ha van 3 példányunk, mindháromnak adhatunk 3 CPU-t. Azonban ha mindegyik terhelés alatt van, akkor a példány számára maximum allokált CPU szám osztva összes allokált CPU számmal arányban részesül a processzorból, ami a példában 33,33%, azaz 1,33 CPU. Input Output Resource Manager (IORM):Nem csak a processzorok használatát szabályozhatjuk, lehetoség van a megosztott storage eroforrásainak felosztására is. Az Input Output Resorce Manager (IORM) alkalmazásával storage szinten tudjuk szabályozni az adatbázisok közötti és azokon belüli minimális I/O szinteket. Database Vault: Ugyanazon adatbázisba konszolidált alkalmazások esetén a rendszergazda szerepkörök szeparálása lehetséges az Oracle Database Vault technológiával. Ezzel elérheto az, hogy biztonságosan konszolidáljuk adatbázisainkat úgy, hogy minden adminisztrátor csak a hozzá tartozó adatokat, objektumokat lássa, módosíthassa.

    Read the article

  • Does NHibernate LINQ support ToLower() in Where() clauses?

    - by Daniel T.
    I have an entity and its mapping: public class Test { public virtual int Id { get; set; } public virtual string Name { get; set; } public virtual string Description { get; set; } } public class TestMap : EntityMap<Test> { public TestMap() { Id(x => x.Id); Map(x => x.Name); Map(x => x.Description); } } I'm trying to run a query on it (to grab it out of the database): var keyword = "test" // this is coming in from the user keyword = keyword.ToLower(); // convert it to all lower-case var results = session.Linq<Test> .Where(x => x.Name.ToLower().Contains(keyword)); results.Count(); // execute the query However, whenever I run this query, I get the following exception: Index was out of range. Must be non-negative and less than the size of the collection. Parameter name: index Am I right when I say that, currently, Linq to NHibernate does not support ToLower()? And if so, is there an alternative that allows me to search for a string in the middle of another string that Linq to NHibernate is compatible with? For example, if the user searches for kap, I need it to match Kapiolani, Makapuu, and Lapkap.

    Read the article

  • Problems configuring nameserver in plesk

    - by Saif Bechan
    Hello, i have some troubles with setting up a nameserver in PLESK for months now. I have tried all possible scenario's but i can not get this to work. I am really in need for some help, and if you can i will really appreciate it. Basically what i want is to just set up a nameserver in PLESK. I have a primary IP, and my host gave me a secondary nameserver i can use. My host is leaseweb in the netherlands. I have made some screenshots of the important parts in my opinion, maybe you guys can see some errors in them. To use the secondary nameserver provided by leaseweb i had to enable ACL on that account, i did so and made a screenshot of that too. The DNS recursion is set to localnets. These settings have not changed for months, so the dns should be fully updated everywhere. The check i run is the following: https://www.sidn.nl/over-nl/aanvraag...-server-check/ Domeinnaam (inclusief .nl): rdshosting.nl Eerste Nameserver: ns1.rdshosting.nl Eerste IP: 62.212.66.33 Tweede Nameserver: ns7.leaseweb.net Tweede ip: 62.212.76.50 If i run the dns check of the netherlands it gives me the following errors: primary name server "ns1.rdshosting.nl." Error: specified name server is not listed as NS record. All public name servers for a domain must also be listed as NS records in the zone of the domain. This domain was specified explicitly as a name server, but not found in the zone description of the primary name server. TE.6a rdshosting.nl. 86400 IN SOA ns1.rdspartners.nl. saif2k.hotmail.com. (2010031102 12H 1H 7D 3H) Error: the MNAME in SOA says "ns1.rdspartners.nl." is the primary name server. The MNAME field in the SOA record (first parameter) lists a different primary name server from the one specified for this check. RFC1035 section 3.3.13 rdshosting.nl. 86400 IN NS ns1.rdspartners.nl. Warning: hidden name server "ns1.rdspartners.nl." never used for first contact. The zone contains an NS record for a host which is not in the list of specified name servers. Hence, this name server will not be used to initiate contact to the domain. It may be used in sequential lookups, so it may still be useful. secondary name server "ns1.rdspartners.nl." [BROKEN] [HIDDEN] Failure: name server at 77.232.85.129 cannot be reached: (unknown error) The name server could not be contacted, which may be due to temporary technical problems or global DNS configuration mistakes. The internal error is shown, but not always clear about the cause. secondary name server "ns7.leaseweb.net." Info: name server looks correctly configured. I have the content of the file etc/named.conf also: // $Id: named.conf,v 1.1.1.1 2001/10/15 07:44:36 kap Exp $ // // Refer to the named(8) man page for details. If you are ever going // to setup a primary server, make sure you've understood the hairy // details of how DNS is working. Even with simple mistakes, you can // break connectivity for affected parties, or cause huge amount of // useless Internet traffic. options { allow-recursion { localnets; }; directory "/var"; auth-nxdomain no; pid-file "/var/run/named/named.pid"; // In addition to the "forwarders" clause, you can force your name // server to never initiate queries of its own, but always ask its // forwarders only, by enabling the following line: // // forward only; // If you've got a DNS server around at your upstream provider, enter // its IP address here, and enable the line below. This will make you // benefit from its cache, thus reduce overall DNS traffic in the Internet. /* forwarders { 127.0.0.1; }; */ /* * If there is a firewall between you and nameservers you want * to talk to, you might need to uncomment the query-source * directive below. Previous versions of BIND always asked * questions using port 53, but BIND 8.1 uses an unprivileged * port by default. */ // query-source address * port 53; /* * If running in a sandbox, you may have to specify a different * location for the dumpfile. */ // dump-file "s/named_dump.db"; }; //Use with the following in named.conf, adjusting the allow list as needed: key "rndc-key" { algorithm hmac-md5; secret "CeMgS23y0oWE20nyv0x40Q=="; }; controls { inet 127.0.0.1 port 953 allow { 127.0.0.1; } keys { "rndc-key"; }; }; // Note: the following will be supported in a future release. /* host { any; } { topology { 127.0.0.0/8; }; }; */ // Setting up secondaries is way easier and the rough picture for this // is explained below. // // If you enable a local name server, don't forget to enter 127.0.0.1 // into your /etc/resolv.conf so this server will be queried first. // Also, make sure to enable it in /etc/rc.conf. zone "." { type hint; file "named.root"; }; zone "0.0.127.IN-ADDR.ARPA" { type master; file "localhost.rev"; }; // NB: Do not use the IP addresses below, they are faked, and only // serve demonstration/documentation purposes! // // Example secondary config entries. It can be convenient to become // a secondary at least for the zone where your own domain is in. Ask // your network administrator for the IP address of the responsible // primary. // // Never forget to include the reverse lookup (IN-ADDR.ARPA) zone! // (This is the first bytes of the respective IP address, in reverse // order, with ".IN-ADDR.ARPA" appended.) // // Before starting to setup a primary zone, better make sure you fully // understand how DNS and BIND works, however. There are sometimes // unobvious pitfalls. Setting up a secondary is comparably simpler. // // NB: Don't blindly enable the examples below. :-) Use actual names // and addresses instead. // // NOTE!!! FreeBSD runs bind in a sandbox (see named_flags in rc.conf). // The directory containing the secondary zones must be write accessible // to bind. The following sequence is suggested: // // mkdir /etc/namedb/s // chown bind.bind /etc/namedb/s // chmod 750 /etc/namedb/s zone "rdshosting.nl" { type master; file "rdshosting.nl"; allow-transfer { 77.232.85.129; 62.212.76.50; common-allow-transfer; }; }; zone "66.212.62.in-addr.arpa" { type master; file "66.212.62.in-addr.arpa"; allow-transfer { common-allow-transfer; }; }; acl common-allow-transfer { 62.212.76.50; }; As i mentioned i made some screenshots of some parts: First the dns settings in plesk: http://www.freeimagehosting.net/uploads/2480faed5e.jpg Second the acl settings in plesk: http://www.freeimagehosting.net/uploads/777f5e69b0.jpg Third my settings at leaseweb: http://www.freeimagehosting.net/uploads/de7122b19c.jpg And last the secondary nameserver settings from leaseweb: http://www.freeimagehosting.net/uploads/fd1da38a8f.jpg If someone has anysuggestion at all on this this will be highly appriciated. Thank you for your time! PS. I am dutch so dutch answers are welcome aswell

    Read the article

1