当我在顶部循环上有一个内核时,为什么我不能使用这两个指令:
#pragma acc update device(hbias[0:n_hidden],W[0:n_hidden][0:n_visible])
#pragma acc update device(vbias[0:n_visible)
我需要在下面的代码中更新这些变量hbias
、vbias
、W
,但它不起作用:
void RBM::contrastive_divergence(int train_X[6][6], double learning_rate, int k) {
double r= rand() / (RAND_MAX + 1.0);
int * input = new int[n_visible];
double *ph_mean = new double[n_hidden];
int *ph_sample = new int[n_hidden];
double *nv_means = new double[n_visible];
int *nv_samples = new int[n_visible];
double *nh_means = new double[n_hidden];
int *nh_samples = new int[n_hidden];
#pragma acc kernels
for (int i = 0; i<train_N; i++) {
for (int j = 0; j< n_visible; j++){
input[j] = train_X[i][j];
}
sample_h_given_v(input, ph_mean, ph_sample,r);
for (int step = 0; step<k; step++) {
if (step == 0) {
gibbs_hvh(ph_sample, nv_means, nv_samples, nh_means, nh_samples,r);
}
else {
gibbs_hvh(nh_samples, nv_means, nv_samples, nh_means, nh_samples,r);
}
}
for (int i = 0; i<n_hidden; i++) {
for (int j = 0; j<n_visible; j++) {
W[i][j] += learning_rate * (ph_mean[i] * input[j] - nh_means[i] * nv_samples[j]) / N;
}
hbias[i] += learning_rate * (ph_sample[i] - nh_means[i]) / N;
}
//this directive
#pragma acc update device(hbias[0:n_hidden],W[0:n_hidden][0:n_visible])
for (int i = 0; i<n_visible; i++) {
vbias[i] += learning_rate * (input[i] - nv_samples[i]) / N;
}
//and this directive
#pragma acc update device(vbias[0:n_visible)
}
delete[] input;
delete[] ph_mean;
delete[] ph_sample;
delete[] nv_means;
delete[] nv_samples;
delete[] nh_means;
delete[] nh_samples;
}
但是当我有许多独立的内核在每个嵌套循环上工作时,我可以更新变量:
void RBM::contrastive_divergence(int train_X[6][6], double learning_rate, int k) {
double r= rand() / (RAND_MAX + 1.0);
int * input = new int[n_visible];
double *ph_mean = new double[n_hidden];
int *ph_sample = new int[n_hidden];
double *nv_means = new double[n_visible];
int *nv_samples = new int[n_visible];
double *nh_means = new double[n_hidden];
int *nh_samples = new int[n_hidden];
for (int i = 0; i<train_N; i++) {
#pragma acc kernels
for (int j = 0; j< n_visible; j++){
input[j] = train_X[i][j];
}
sample_h_given_v(input, ph_mean, ph_sample,r);
#pragma acc kernels
for (int step = 0; step<k; step++) {
if (step == 0) {
gibbs_hvh(ph_sample, nv_means, nv_samples, nh_means, nh_samples,r);
}
else {
gibbs_hvh(nh_samples, nv_means, nv_samples, nh_means, nh_samples,r);
}
}
#pragma acc kernels
{
for (int i = 0; i<unhidden; i++) {
for (int j = 0; j<n_visible; j++) {
W[i][j] += learning_rate * (ph_mean[i] * input[j] - nh_means[i] * nv_samples[j]) / N;
}
hbias[i] += learning_rate * (ph_sample[i] - nh_means[i]) / N;
}
//this directive
#pragma acc update device(hbias[0:n_hidden],W[0:n_hidden][0:n_visible])
}
#pragma acc kernels
{
for (int i = 0; i<n_visible; i++) {
vbias[i] += learning_rate * (input[i] - nv_samples[i]) / N;
}
//and this directive
#pragma acc update device(vbias[0:n_visible)
}
}
delete[] input;
delete[] ph_mean;
delete[] ph_sample;
delete[] nv_means;
delete[] nv_samples;
delete[] nh_means;
delete[] nh_samples;
}
最佳答案
“更新”指令只能在主机代码中使用,因为数据移动必须从主机发起。您不能将它们放在计算区域内。
这段代码有很多问题。首先,对于嵌套循环使用相同的索引变量(在本例中为“i”)可能是不好的做法。尽管范围规则允许这样做,但很难判断代码应该使用哪个“i”。
外部“i”循环可能不安全地并行化,因此您不应该将“kernels”指令放在该循环之外。也许如果你私有(private)化“输入”数组,然后在更新 vbias、hbias、W 数组时使用原子,它可能会起作用,但你的性能会很差。 (您还需要确定其他数组是否需要私有(private)化或者是全局的,因此需要原子操作)。
我建议首先在内部循环周围放置“#pragma accparallelloop”,一次一个。确保每一项都能正常工作,然后再继续下一项。另外,我非常怀疑“step”循环是否可并行化,因此您很可能需要并行化“gibbs_hvh”子例程内的循环。
由于您使用的是 CUDA 统一内存 (-ta=tesla: Managed),因此可能不需要添加数据区域。但是,如果您计划将来不使用托管内存,下一步将是在外部“i”循环周围添加数据指令(或在程序的更高点,然后使用更新指令来同步外部“i”循环之后的数据)我”循环)。
关于c++ - 更新指令 OpenACC,我们在Stack Overflow上找到一个类似的问题: https://stackoverflow.com/questions/41665671/