Skip to content

[llama32_1b] 3-launch o_gemv_ffn for decode (+17% tok/s) #1631

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
erwei-xilinx merged 3 commits into from
May 30, 2026
+1,829 −578
Merged

[llama32_1b] 3-launch o_gemv_ffn for decode (+17% tok/s) #1631

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
c04003e
[llama32_1b] 3-launch o_gemv_ffn for decode
erwei-xilinx May 28, 2026
64c11e8
Merge branch 'main' into decode-3launch-o-gemv-ffn
erwei-xilinx May 29, 2026
c4b8839
[llama32_1b] address Copilot review comments on #1631
erwei-xilinx May 29, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
40 changes: 40 additions & 0 deletions programming_examples/decode_ffn_swiglu/Makefile
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,40 @@
# Copyright (C) 2026, Advanced Micro Devices, Inc.
# SPDX-License-Identifier: MIT
#
# Single-token GEMV with fused weighted RMSNorm input and SwiGLU output.
srcdir := $(shell dirname $(realpath $(firstword $(MAKEFILE_LIST))))

ifdef PEANO_INSTALL_DIR
BUILD_DIR := build_peano
else
BUILD_DIR := build_chess
endif

OUTPUT_FORMAT ?= xclbin
OUTPUT_FORMAT_FLAG = --output-format $(OUTPUT_FORMAT)

# Default shape: M = 2*hidden, K = emb_dim. Tuned for an 8-col herd at
# emb_dim=2048 / hidden_dim=8192.
M ?= 16384
K ?= 2048
TILE_M ?= 32
M_INPUT ?= 4
HERD_COLS ?= 8
N_CASCADE ?= 4

all: run

print:
${powershell} python3 ${srcdir}/matvec_swiglu_rms.py $(OUTPUT_FORMAT_FLAG) -p \
--m $(M) --k $(K) --tile-m $(TILE_M) --m-input $(M_INPUT) \
--herd-cols $(HERD_COLS) --n-cascade $(N_CASCADE)

run:
mkdir -p $(BUILD_DIR)
PEANO_INSTALL_DIR=$(PEANO_INSTALL_DIR) cd $(BUILD_DIR) && \
${powershell} python3 ${srcdir}/matvec_swiglu_rms.py $(OUTPUT_FORMAT_FLAG) \
--m $(M) --k $(K) --tile-m $(TILE_M) --m-input $(M_INPUT) \
--herd-cols $(HERD_COLS) --n-cascade $(N_CASCADE)

clean:
rm -rf $(BUILD_DIR) __pycache__
Loading
Loading