diff options
| author | snadampal <87143774+snadampal@users.noreply.github.com> | 2024-02-11 07:22:33 -0600 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2024-02-11 15:22:33 +0200 |
| commit | a07d0fee1f05c5c1dc49948ae1a3293db017275f (patch) | |
| tree | 06614ff1364269493e4853333ced56802abd7284 /pocs | |
| parent | e4640d8fdf56f14a6db3d092bcd3d2d315cb5d04 (diff) | |
ggml : add mmla kernels for quantized GEMM (#4966)
* ggml: aarch64: implement smmla kernel for q8_0_q8_0 quantized gemm
armv8.2-a and above supports MMLA instructions that have higher
throughput than DOT. this commit adds mmla kernel for
q8_0_q8_0 gemm. The feature is enabled if the platform supports
"__ARM_FEATURE_MATMUL_INT8"
On AWS Graviton3 processors this kernel resulted up to 1.5x
improvement for prompt evaluation throughput compared to the
default sdot kernel.
* ggml: aarch64: implement smmla kernel for q4_0_q8_0 quantized gemm
armv8.2-a and above supports MMLA instructions that have higher
throughput than DOT. this commit adds mmla kernel for
q4_0_q8_0 gemm. The feature is enabled if the platform supports
"__ARM_FEATURE_MATMUL_INT8"
On AWS Graviton3 processors this kernel resulted up to 1.5x
improvement for prompt evaluation throughput compared to the
default sdot kernel.
* ggml: aarch64: implement smmla kernel for q4_1_q8_1 quantized gemm
armv8.2-a and above supports MMLA instructions that have higher
throughput than DOT. this commit adds mmla kernel for
q4_1_q8_1 gemm. The feature is enabled if the platform supports
"__ARM_FEATURE_MATMUL_INT8"
On AWS Graviton3 processors this kernel resulted up to 1.5x
improvement for prompt evaluation throughput compared to the
default sdot kernel.
* ggml: update unit tests for the new vec_dot interface
* llama.cpp: add MATMUL_INT8 capability to system_info
Diffstat (limited to 'pocs')
| -rw-r--r-- | pocs/vdot/q8dot.cpp | 4 | ||||
| -rw-r--r-- | pocs/vdot/vdot.cpp | 4 |
2 files changed, 4 insertions, 4 deletions
diff --git a/pocs/vdot/q8dot.cpp b/pocs/vdot/q8dot.cpp index 111770d5..1a52ff5e 100644 --- a/pocs/vdot/q8dot.cpp +++ b/pocs/vdot/q8dot.cpp @@ -156,8 +156,8 @@ int main(int argc, char** argv) { t1 = std::chrono::high_resolution_clock::now(); float fs; - if (type == 0) funcs.vec_dot(kVecSize * QK4_1, &fs, x40.data(), y.data()); - else funcs.vec_dot(kVecSize * QK4_1, &fs, x41.data(), y.data()); + if (type == 0) funcs.vec_dot(kVecSize * QK4_1, &fs, 0, x40.data(), 0, y.data(), 0, 1); + else funcs.vec_dot(kVecSize * QK4_1, &fs, 0, x41.data(), 0, y.data(), 0, 1); t2 = std::chrono::high_resolution_clock::now(); t = 1e-3*std::chrono::duration_cast<std::chrono::nanoseconds>(t2-t1).count(); if (iloop > 3) ggml.addResult(fs, t); diff --git a/pocs/vdot/vdot.cpp b/pocs/vdot/vdot.cpp index 73ffcd1c..17e9e448 100644 --- a/pocs/vdot/vdot.cpp +++ b/pocs/vdot/vdot.cpp @@ -284,8 +284,8 @@ int main(int argc, char** argv) { else { auto vdot = ggml_internal_get_type_traits(funcs.vec_dot_type); vdot.from_float(y1.data(), q8.data(), kVecSize); - if (useQ4_1) funcs.vec_dot(kVecSize, &result, q41.data(), q8.data()); - else funcs.vec_dot(kVecSize, &result, q40.data(), q8.data()); + if (useQ4_1) funcs.vec_dot(kVecSize, &result, 0, q41.data(), 0, q8.data(), 0, 1); + else funcs.vec_dot(kVecSize, &result, 0, q40.data(), 0, q8.data(), 0, 1); } sumq += result; t2 = std::chrono::high_resolution_clock::now(); |