ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-12 09:16:52 +00:00
Code Issues Packages Projects Releases Wiki Activity

Add last steps: activations functions

Browse Source
This commit is contained in:
Damien Lejeune
2026-01-29 08:30:45 -05:00
parent da895cdd88
commit 6ea40157f1
4 changed files with 190 additions and 26 deletions
Download Patch File Download Diff File Expand all files Collapse all files

31
include/ck_tile/host/reference/reference_mhc.hpp
View File

@@ -7,6 +7,7 @@
#include <thread>
#include "ck_tile/core.hpp"
#include "ck_tile/host/host_tensor.hpp"
#include "ck_tile/ops/elementwise/unary_element_wise_operation.hpp"
namespace ck_tile {
@@ -14,17 +15,19 @@ namespace ck_tile {
template <typename XDataType,
typename PhiDataType,
typename YDataType,
typename ComputeDataType = float>
typename ComputeDataType = float,
typename Activation = element_wise::Sigmoid>
CK_TILE_HOST void reference_mhc(const HostTensor<XDataType>& x_b_nc, // [B, nC]
const HostTensor<PhiDataType>& phi_nc_out, // [nC, 2n+n²]
HostTensor<YDataType>& output_b_out, // [B, 2n+n²]
const HostTensor<PhiDataType>& phi_nc_out, // [nC, 2n+n^2]
HostTensor<YDataType>& output_b_out, // [B, 2n+n^2]
int n, // expansion factor
int C, // channels per stream
[[maybe_unused]] float r = 1.0f,
[[maybe_unused]] float alpha_pre = 1.0f,
[[maybe_unused]] float alpha_post = 1.0f,
[[maybe_unused]] float alpha_res = 1.0f,
[[maybe_unused]] float bias = 0.0f)
[[maybe_unused]] float bias = 0.0f,
Activation activation = Activation{})
{
const int B = x_b_nc.get_length(0);
const int nC = n * C;
@@ -43,7 +46,7 @@ CK_TILE_HOST void reference_mhc(const HostTensor<XDataType>& x_b_nc, // [B
// Step 2 & 3: Perform GEMM and apply elementwise operations
// Process H^{pre}: x * phi[:, 0:n] -> output[:, 0:n]
// Process H^{pre}: x * phi[:, 0:n] -> sigma(output[:, 0:n])
for(int out_idx = 0; out_idx < n; out_idx++)
{
ComputeDataType sum = 0.0f;
@@ -52,11 +55,14 @@ CK_TILE_HOST void reference_mhc(const HostTensor<XDataType>& x_b_nc, // [B
sum += type_convert<ComputeDataType>(x_b_nc(b, k)) *
type_convert<ComputeDataType>(phi_nc_out(k, out_idx));
}
// Apply: 1/r * alpha_pre * sum + bias
output_b_out(b, out_idx) = type_convert<YDataType>((alpha_pre / r) * sum + bias);
// Step 4: Apply activation σ(H^{pre})
ComputeDataType activated_value;
activation(activated_value, sum);
output_b_out(b, out_idx) =
type_convert<YDataType>((alpha_pre / r) * activated_value + bias);
}
// Process H^{post}: x * phi[:, n:2n] -> output[:, n:2n]
// Process H^{post}: x * phi[:, n:2n] -> 2*sigma(output[:, n:2n])
for(int out_idx = 0; out_idx < n; out_idx++)
{
ComputeDataType sum = 0.0f;
@@ -65,11 +71,14 @@ CK_TILE_HOST void reference_mhc(const HostTensor<XDataType>& x_b_nc, // [B
sum += type_convert<ComputeDataType>(x_b_nc(b, k)) *
type_convert<ComputeDataType>(phi_nc_out(k, n + out_idx));
}
// Apply: 1/r * alpha_post * sum + bias
output_b_out(b, n + out_idx) = type_convert<YDataType>((alpha_post / r) * sum + bias);
// Step 5: Apply 2*σ(H^{post})
ComputeDataType activated_value;
activation(activated_value, sum);
output_b_out(b, n + out_idx) =
type_convert<YDataType>((alpha_post / r) * 2.0f * activated_value + bias);
}
// Process H^{res}: x * phi[:, 2n:2n+n²] -> output[:, 2n:2n+n²]
// Process H^{res}: x * phi[:, 2n:2n+n^2] -> output[:, 2n:2n+n^2]
int n_squared = n * n;
for(int out_idx = 0; out_idx < n_squared; out_idx++)
{

49
include/ck_tile/ops/mhc/kernel/mhc_kernel_tile_v2.hpp
View File

@@ -8,6 +8,7 @@
#include "ck_tile/ops/mhc/pipeline/mhc_problem.hpp"
#include "ck_tile/ops/mhc/pipeline/mhc_default_policy.hpp"
#include "ck_tile/ops/gemm/block/block_gemm_asmem_bsmem_creg_v1.hpp"
#include "ck_tile/ops/elementwise/unary_element_wise_operation.hpp"
// Manifold Constrained Hyper Connection Kernel (True CK Tile Version):
// =====================================================================
@@ -20,15 +21,17 @@
namespace ck_tile {
template <typename Problem_,
typename Policy_ = MHCDefaultPolicy,
index_t B_ = 16, // Batch size (compile-time)
index_t N_ = 4, // Expansion factor (compile-time)
index_t C_ = 64, // Channels per stream (compile-time)
index_t KTile_ = 256> // K-tile size for shared memory (compile-time)
typename Policy_ = MHCDefaultPolicy,
index_t B_ = 16, // Batch size (compile-time)
index_t N_ = 4, // Expansion factor (compile-time)
index_t C_ = 64, // Channels per stream (compile-time)
index_t KTile_ = 256, // K-tile size for shared memory (compile-time)
typename Activation_ = element_wise::Sigmoid> // Activation function (compile-time)
struct ManifoldConstrainedHyperConnectionTiled
{
using Problem = ck_tile::remove_cvref_t<Problem_>;
using Policy = ck_tile::remove_cvref_t<Policy_>;
using Activation = ck_tile::remove_cvref_t<Activation_>;
using Problem = ck_tile::remove_cvref_t<Problem_>;
using Policy = ck_tile::remove_cvref_t<Policy_>;
using XDataType = ck_tile::remove_cvref_t<typename Problem::XDataType>;
using ComputeDataType = ck_tile::remove_cvref_t<typename Problem::ComputeDataType>;
@@ -241,16 +244,32 @@ struct ManifoldConstrainedHyperConnectionTiled
if(global_batch < B && global_col < output_dim)
{
// Determine alpha based on the actual output column
float alpha = (global_col < kN) ? alpha_pre
: (global_col < 2 * kN) ? alpha_post
: alpha_res;
constexpr auto i_j_idx = make_tuple(idx0, idx1);
ComputeDataType value = result_tile[i_j_idx];
// Step 4 & 5: Apply activation functions based on output section
if(global_col < kN)
{
// H^{pre}: Apply sigma(H^{pre})
ComputeDataType activated_value;
Activation{}(activated_value, value);
value = (alpha_pre / r) * activated_value + bias;
}
else if(global_col < 2 * kN)
{
// H^{post}: Apply 2*sigma(H^{post})
ComputeDataType activated_value;
Activation{}(activated_value, value);
value = (alpha_post / r) * 2.0f * activated_value + bias;
}
else
{
// H^{res}: No activation (will be Sinkhorn-Knopp later)
value = (alpha_res / r) * value + bias;
}
// Apply scaling and bias, then store: result = (alpha / r) * result + bias
constexpr auto i_j_idx = make_tuple(idx0, idx1);
const index_t global_idx = global_batch * output_dim + global_col;
p_output[global_idx] =
type_convert<YDataType>((alpha / r) * result_tile[i_j_idx] + bias);
p_output[global_idx] = type_convert<YDataType>(value);
}
});
});