Skip to content

Commit

Permalink
skip more halfway cases
Browse files Browse the repository at this point in the history
  • Loading branch information
@nihui
nihui committed Dec 12, 2024
1 parent 6347dd1 commit 82f1a2b
Show file tree
Hide file tree
Showing 2 changed files with 196 additions and 50 deletions.
127 changes: 100 additions & 27 deletions tests/test_gemm_3.cpp
View file
Original file line number Diff line number Diff line change
Expand Up @@ -23,7 +23,10 @@ static void RandomizeA(ncnn::Mat& m, int transA, float absmax)
for (int i = 0; i < h; i++)
{
float* p = m.dims == 3 ? m.channel(i) : m.row(i);
const float randabsmax = RandomFloat(absmax * 0.5f, absmax);
float randabsmax = RandomFloat(absmax * 0.5f, absmax);
randabsmax = ncnn::float16_to_float32(ncnn::float32_to_float16(randabsmax));
randabsmax = ncnn::bfloat16_to_float32(ncnn::float32_to_bfloat16(randabsmax));

for (int j = 0; j < m.w; j++)
{
p[j] = RandomFloat(-randabsmax, randabsmax);
Expand All @@ -33,16 +36,33 @@ static void RandomizeA(ncnn::Mat& m, int transA, float absmax)
p[RandomInt(0, m.w - 1)] = -randabsmax;
p[RandomInt(0, m.w - 1)] = randabsmax;

// drop 0.4 ~ 0.6
// drop 0.45 ~ 0.55
for (int j = 0; j < m.w; j++)
{
float v = p[j] / randabsmax * 127.f;
float v = p[j] * (127.f / randabsmax);
float vv = fabs(v - (int)v);
while (vv > 0.4f && vv < 0.6f)

float hp = ncnn::float16_to_float32(ncnn::float32_to_float16(p[j]));
float hv = hp * (127.f / randabsmax);
float hvv = fabs(hv - (int)hv);

float bp = ncnn::bfloat16_to_float32(ncnn::float32_to_bfloat16(p[j]));
float bv = bp * (127.f / randabsmax);
float bvv = fabs(bv - (int)bv);

while ((vv > 0.45f && vv < 0.55f) || (hvv > 0.45f && hvv < 0.55f) || (bvv > 0.45f && bvv < 0.55f))
{
p[j] = RandomFloat(-randabsmax, randabsmax);
v = p[j] / randabsmax * 127.f;
v = p[j] * (127.f / randabsmax);
vv = fabs(v - (int)v);

hp = ncnn::float16_to_float32(ncnn::float32_to_float16(p[j]));
hv = hp * (127.f / randabsmax);
hvv = fabs(hv - (int)hv);

bp = ncnn::bfloat16_to_float32(ncnn::float32_to_bfloat16(p[j]));
bv = bp * (127.f / randabsmax);
bvv = fabs(bv - (int)bv);
}
}
}
Expand All @@ -52,7 +72,10 @@ static void RandomizeA(ncnn::Mat& m, int transA, float absmax)
std::vector<float> randabsmaxes(m.w);
for (int j = 0; j < m.w; j++)
{
randabsmaxes[j] = RandomFloat(absmax * 0.5f, absmax);
float randabsmax = RandomFloat(absmax * 0.5f, absmax);
randabsmax = ncnn::float16_to_float32(ncnn::float32_to_float16(randabsmax));
randabsmax = ncnn::bfloat16_to_float32(ncnn::float32_to_bfloat16(randabsmax));
randabsmaxes[j] = randabsmax;
}

const int h = m.dims == 3 ? m.c : m.h;
Expand All @@ -65,17 +88,34 @@ static void RandomizeA(ncnn::Mat& m, int transA, float absmax)
p[j] = RandomFloat(-randabsmax, randabsmax);
}

// drop 0.4 ~ 0.6
// drop 0.45 ~ 0.55
for (int j = 0; j < m.w; j++)
{
const float randabsmax = randabsmaxes[j];
float v = p[j] / randabsmax * 127.f;
float v = p[j] * (127.f / randabsmax);
float vv = fabs(v - (int)v);
while (vv > 0.4f && vv < 0.6f)

float hp = ncnn::float16_to_float32(ncnn::float32_to_float16(p[j]));
float hv = hp * (127.f / randabsmax);
float hvv = fabs(hv - (int)hv);

float bp = ncnn::bfloat16_to_float32(ncnn::float32_to_bfloat16(p[j]));
float bv = bp * (127.f / randabsmax);
float bvv = fabs(bv - (int)bv);

while ((vv > 0.45f && vv < 0.55f) || (hvv > 0.45f && hvv < 0.55f) || (bvv > 0.45f && bvv < 0.55f))
{
p[j] = RandomFloat(-randabsmax, randabsmax);
v = p[j] / randabsmax * 127.f;
v = p[j] * (127.f / randabsmax);
vv = fabs(v - (int)v);

hp = ncnn::float16_to_float32(ncnn::float32_to_float16(p[j]));
hv = hp * (127.f / randabsmax);
hvv = fabs(hv - (int)hv);

bp = ncnn::bfloat16_to_float32(ncnn::float32_to_bfloat16(p[j]));
bv = bp * (127.f / randabsmax);
bvv = fabs(bv - (int)bv);
}
}
}
Expand All @@ -98,25 +138,58 @@ static void RandomizeA(ncnn::Mat& m, int transA, float absmax)

static void RandomizeB(ncnn::Mat& m, float absmax)
{
absmax = ncnn::float16_to_float32(ncnn::float32_to_float16(absmax));
absmax = ncnn::bfloat16_to_float32(ncnn::float32_to_bfloat16(absmax));

const int h = m.dims == 3 ? m.c : m.h;
float* p = m;
for (int i = 0; i < m.total(); i++)
for (int i = 0; i < h; i++)
{
p[i] = RandomFloat(-absmax, absmax);
float* p = m.dims == 3 ? m.channel(i) : m.row(i);
for (int j = 0; j < m.w; j++)
{
p[j] = RandomFloat(-absmax, absmax);

// set random a and b
p[RandomInt(0, m.total() - 1)] = -absmax;
p[RandomInt(0, m.total() - 1)] = absmax;
// drop 0.45 ~ 0.55
float v = p[j] * (127.f / absmax);
float vv = fabs(v - (int)v);

// drop 0.4 ~ 0.6
float v = p[i] / absmax * 127.f;
float vv = fabs(v - (int)v);
while (vv > 0.4f && vv < 0.6f)
{
p[i] = RandomFloat(-absmax, absmax);
v = p[i] / absmax * 127.f;
vv = fabs(v - (int)v);
float hp = ncnn::float16_to_float32(ncnn::float32_to_float16(p[j]));
float hv = hp * (127.f / absmax);
float hvv = fabs(hv - (int)hv);

float bp = ncnn::bfloat16_to_float32(ncnn::float32_to_bfloat16(p[j]));
float bv = bp * (127.f / absmax);
float bvv = fabs(bv - (int)bv);

while ((vv > 0.45f && vv < 0.55f) || (hvv > 0.45f && hvv < 0.55f) || (bvv > 0.45f && bvv < 0.55f))
{
p[j] = RandomFloat(-absmax, absmax);
v = p[j] * (127.f / absmax);
vv = fabs(v - (int)v);

hp = ncnn::float16_to_float32(ncnn::float32_to_float16(p[j]));
hv = hp * (127.f / absmax);
hvv = fabs(hv - (int)hv);

bp = ncnn::bfloat16_to_float32(ncnn::float32_to_bfloat16(p[j]));
bv = bp * (127.f / absmax);
bvv = fabs(bv - (int)bv);
}
}
}

// set random a and b
if (m.dims == 3)
{
m.channel(RandomInt(0, h - 1))[RandomInt(0, m.w - 1)] = -absmax;
m.channel(RandomInt(0, h - 1))[RandomInt(0, m.w - 1)] = absmax;
}
else
{
m.row(RandomInt(0, h - 1))[RandomInt(0, m.w - 1)] = -absmax;
m.row(RandomInt(0, h - 1))[RandomInt(0, m.w - 1)] = absmax;
}
}

static int test_gemm_int8(int M, int N, int K, float alpha, int transA, int transB, int output_elemtype, int output_transpose, int constantA, int constantB, int output_N1M)
Expand All @@ -139,8 +212,8 @@ static int test_gemm_int8(int M, int N, int K, float alpha, int transA, int tran
pd.set(18, 2); // int8_scale_term

std::vector<ncnn::Mat> weights;
if (constantA) weights.push_back(transA ? (output_N1M ? RandomS8Mat(M, 1, K) : RandomS8Mat(M, K)) : (output_N1M ? RandomS8Mat(K, 1, M) : RandomS8Mat(K, M)));
if (constantB) weights.push_back(transB ? (output_N1M ? RandomS8Mat(K, 1, N) : RandomS8Mat(K, N)) : (output_N1M ? RandomS8Mat(N, 1, K) : RandomS8Mat(N, K)));
if (constantA) weights.push_back(transA ? RandomS8Mat(M, K) : RandomS8Mat(K, M));
if (constantB) weights.push_back(transB ? RandomS8Mat(K, N) : RandomS8Mat(N, K));
if (constantA) weights.push_back(RandomMat(M, 10.f, 20.f));
if (constantB) weights.push_back(RandomMat(1, 10.f, 20.f));

Expand Down Expand Up @@ -266,8 +339,8 @@ static int test_gemm_int8_fp16s(int M, int N, int K, float alpha, int transA, in
pd.set(18, 2); // int8_scale_term

std::vector<ncnn::Mat> weights;
if (constantA) weights.push_back(transA ? (output_N1M ? RandomS8Mat(M, 1, K) : RandomS8Mat(M, K)) : (output_N1M ? RandomS8Mat(K, 1, M) : RandomS8Mat(K, M)));
if (constantB) weights.push_back(transB ? (output_N1M ? RandomS8Mat(K, 1, N) : RandomS8Mat(K, N)) : (output_N1M ? RandomS8Mat(N, 1, K) : RandomS8Mat(N, K)));
if (constantA) weights.push_back(transA ? RandomS8Mat(M, K) : RandomS8Mat(K, M));
if (constantB) weights.push_back(transB ? RandomS8Mat(K, N) : RandomS8Mat(N, K));
if (constantA) weights.push_back(RandomMat(M, 10.f, 20.f));
if (constantB) weights.push_back(RandomMat(1, 10.f, 20.f));

Expand Down
119 changes: 96 additions & 23 deletions tests/test_gemm_4.cpp
View file
Original file line number Diff line number Diff line change
Expand Up @@ -23,7 +23,10 @@ static void RandomizeA(ncnn::Mat& m, int transA, float absmax)
for (int i = 0; i < h; i++)
{
float* p = m.dims == 3 ? m.channel(i) : m.row(i);
const float randabsmax = RandomFloat(absmax * 0.5f, absmax);
float randabsmax = RandomFloat(absmax * 0.5f, absmax);
randabsmax = ncnn::float16_to_float32(ncnn::float32_to_float16(randabsmax));
randabsmax = ncnn::bfloat16_to_float32(ncnn::float32_to_bfloat16(randabsmax));

for (int j = 0; j < m.w; j++)
{
p[j] = RandomFloat(-randabsmax, randabsmax);
Expand All @@ -33,16 +36,33 @@ static void RandomizeA(ncnn::Mat& m, int transA, float absmax)
p[RandomInt(0, m.w - 1)] = -randabsmax;
p[RandomInt(0, m.w - 1)] = randabsmax;

// drop 0.4 ~ 0.6
// drop 0.45 ~ 0.55
for (int j = 0; j < m.w; j++)
{
float v = p[j] / randabsmax * 127.f;
float v = p[j] * (127.f / randabsmax);
float vv = fabs(v - (int)v);
while (vv > 0.4f && vv < 0.6f)

float hp = ncnn::float16_to_float32(ncnn::float32_to_float16(p[j]));
float hv = hp * (127.f / randabsmax);
float hvv = fabs(hv - (int)hv);

float bp = ncnn::bfloat16_to_float32(ncnn::float32_to_bfloat16(p[j]));
float bv = bp * (127.f / randabsmax);
float bvv = fabs(bv - (int)bv);

while ((vv > 0.45f && vv < 0.55f) || (hvv > 0.45f && hvv < 0.55f) || (bvv > 0.45f && bvv < 0.55f))
{
p[j] = RandomFloat(-randabsmax, randabsmax);
v = p[j] / randabsmax * 127.f;
v = p[j] * (127.f / randabsmax);
vv = fabs(v - (int)v);

hp = ncnn::float16_to_float32(ncnn::float32_to_float16(p[j]));
hv = hp * (127.f / randabsmax);
hvv = fabs(hv - (int)hv);

bp = ncnn::bfloat16_to_float32(ncnn::float32_to_bfloat16(p[j]));
bv = bp * (127.f / randabsmax);
bvv = fabs(bv - (int)bv);
}
}
}
Expand All @@ -52,7 +72,10 @@ static void RandomizeA(ncnn::Mat& m, int transA, float absmax)
std::vector<float> randabsmaxes(m.w);
for (int j = 0; j < m.w; j++)
{
randabsmaxes[j] = RandomFloat(absmax * 0.5f, absmax);
float randabsmax = RandomFloat(absmax * 0.5f, absmax);
randabsmax = ncnn::float16_to_float32(ncnn::float32_to_float16(randabsmax));
randabsmax = ncnn::bfloat16_to_float32(ncnn::float32_to_bfloat16(randabsmax));
randabsmaxes[j] = randabsmax;
}

const int h = m.dims == 3 ? m.c : m.h;
Expand All @@ -65,17 +88,34 @@ static void RandomizeA(ncnn::Mat& m, int transA, float absmax)
p[j] = RandomFloat(-randabsmax, randabsmax);
}

// drop 0.4 ~ 0.6
// drop 0.45 ~ 0.55
for (int j = 0; j < m.w; j++)
{
const float randabsmax = randabsmaxes[j];
float v = p[j] / randabsmax * 127.f;
float v = p[j] * (127.f / randabsmax);
float vv = fabs(v - (int)v);
while (vv > 0.4f && vv < 0.6f)

float hp = ncnn::float16_to_float32(ncnn::float32_to_float16(p[j]));
float hv = hp * (127.f / randabsmax);
float hvv = fabs(hv - (int)hv);

float bp = ncnn::bfloat16_to_float32(ncnn::float32_to_bfloat16(p[j]));
float bv = bp * (127.f / randabsmax);
float bvv = fabs(bv - (int)bv);

while ((vv > 0.45f && vv < 0.55f) || (hvv > 0.45f && hvv < 0.55f) || (bvv > 0.45f && bvv < 0.55f))
{
p[j] = RandomFloat(-randabsmax, randabsmax);
v = p[j] / randabsmax * 127.f;
v = p[j] * (127.f / randabsmax);
vv = fabs(v - (int)v);

hp = ncnn::float16_to_float32(ncnn::float32_to_float16(p[j]));
hv = hp * (127.f / randabsmax);
hvv = fabs(hv - (int)hv);

bp = ncnn::bfloat16_to_float32(ncnn::float32_to_bfloat16(p[j]));
bv = bp * (127.f / randabsmax);
bvv = fabs(bv - (int)bv);
}
}
}
Expand All @@ -98,25 +138,58 @@ static void RandomizeA(ncnn::Mat& m, int transA, float absmax)

static void RandomizeB(ncnn::Mat& m, float absmax)
{
absmax = ncnn::float16_to_float32(ncnn::float32_to_float16(absmax));
absmax = ncnn::bfloat16_to_float32(ncnn::float32_to_bfloat16(absmax));

const int h = m.dims == 3 ? m.c : m.h;
float* p = m;
for (int i = 0; i < m.total(); i++)
for (int i = 0; i < h; i++)
{
p[i] = RandomFloat(-absmax, absmax);
float* p = m.dims == 3 ? m.channel(i) : m.row(i);
for (int j = 0; j < m.w; j++)
{
p[j] = RandomFloat(-absmax, absmax);

// set random a and b
p[RandomInt(0, m.total() - 1)] = -absmax;
p[RandomInt(0, m.total() - 1)] = absmax;
// drop 0.45 ~ 0.55
float v = p[j] * (127.f / absmax);
float vv = fabs(v - (int)v);

// drop 0.4 ~ 0.6
float v = p[i] / absmax * 127.f;
float vv = fabs(v - (int)v);
while (vv > 0.4f && vv < 0.6f)
{
p[i] = RandomFloat(-absmax, absmax);
v = p[i] / absmax * 127.f;
vv = fabs(v - (int)v);
float hp = ncnn::float16_to_float32(ncnn::float32_to_float16(p[j]));
float hv = hp * (127.f / absmax);
float hvv = fabs(hv - (int)hv);

float bp = ncnn::bfloat16_to_float32(ncnn::float32_to_bfloat16(p[j]));
float bv = bp * (127.f / absmax);
float bvv = fabs(bv - (int)bv);

while ((vv > 0.45f && vv < 0.55f) || (hvv > 0.45f && hvv < 0.55f) || (bvv > 0.45f && bvv < 0.55f))
{
p[j] = RandomFloat(-absmax, absmax);
v = p[j] * (127.f / absmax);
vv = fabs(v - (int)v);

hp = ncnn::float16_to_float32(ncnn::float32_to_float16(p[j]));
hv = hp * (127.f / absmax);
hvv = fabs(hv - (int)hv);

bp = ncnn::bfloat16_to_float32(ncnn::float32_to_bfloat16(p[j]));
bv = bp * (127.f / absmax);
bvv = fabs(bv - (int)bv);
}
}
}

// set random a and b
if (m.dims == 3)
{
m.channel(RandomInt(0, h - 1))[RandomInt(0, m.w - 1)] = -absmax;
m.channel(RandomInt(0, h - 1))[RandomInt(0, m.w - 1)] = absmax;
}
else
{
m.row(RandomInt(0, h - 1))[RandomInt(0, m.w - 1)] = -absmax;
m.row(RandomInt(0, h - 1))[RandomInt(0, m.w - 1)] = absmax;
}
}

static int test_gemm_int8(int M, int N, int K, int TILE_M, int TILE_N, int TILE_K, float alpha, int transA, int transB, int output_transpose)
Expand Down

0 comments on commit 82f1a2b

Please sign in to comment.