Skip to content

Pull requests: shalinib-ibm/llama.cpp

New pull request New
23 Open 1 Closed
23 Open 1 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

ggml-cpu : Add Power12 MMA+ INT8 matmul kernels
#24 opened May 6, 2026 by shalinib-ibm Owner Loading…
Q8 Matmul MMA+ Kernels Implementation
#23 opened Apr 22, 2026 by shalinib-ibm Owner Loading…
Q4/Q8 Tiled Gemm Optimization. ggml
#22 opened Oct 26, 2025 by shalinib-ibm Owner Loading…
Q4 Tiled Gemm Optimization ggml
#21 opened Oct 23, 2025 by shalinib-ibm Owner Loading…
Prepack all k blocks of Matrices A and B ggml
#20 opened Oct 8, 2025 by shalinib-ibm Owner Loading…
Pre Compute scale factor once and reuse inside kenrel. ggml
#19 opened Oct 8, 2025 by shalinib-ibm Owner Loading…
Implement a 16x8 kernel for Q4 GEMM ggml
#18 opened Sep 24, 2025 by shalinib-ibm Owner Loading…
Q4 tiled Gemm Implementation ggml
#17 opened Sep 18, 2025 by shalinib-ibm Owner Loading…
Q4 gemm trail1 ggml
#16 opened Sep 12, 2025 by shalinib-ibm Owner Loading…
PowerPC: Sgemm Optimization
#15 opened Aug 5, 2025 by shalinib-ibm Owner Loading…
POC:Avoid PackTranspose TinyBLAS_PPC in MMA kernel
#14 opened Jul 15, 2025 by shalinib-ibm Owner Loading…
PowerPC: Replace static_assert with runtime assertion
#13 opened Jul 14, 2025 by shalinib-ibm Owner Loading…
Exp: Perf Benefit with xvi8ger4pp signed version
#12 opened Jun 16, 2025 by shalinib-ibm Owner Loading…
Move packing from inside MMA kernels
#11 opened Jun 9, 2025 by shalinib-ibm Owner Loading…
Refactor: Move matrix packing outside GEMM kernels
#10 opened Jun 5, 2025 by shalinib-ibm Owner Loading…
Avoid PackTranspose calls Fp32 tiny blas kernels
#9 opened May 30, 2025 by shalinib-ibm Owner Loading…
2
Optimize ggml_vec_dot_q8_q8 for ppc
#8 opened May 28, 2025 by shalinib-ibm Owner Loading…
Manual unrolling of loop for vec_dot_q8_q8
#7 opened May 28, 2025 by shalinib-ibm Owner Loading…
1
ggml_vec_dot_q8_q8 mma optimize
#6 opened May 28, 2025 by shalinib-ibm Owner Loading…
Bug fix
#5 opened Apr 29, 2025 by shalinib-ibm Owner Loading…
PowerPC: Enable MMA for BF16 in llamafile_sgemm
#4 opened Apr 28, 2025 by shalinib-ibm Owner Loading…
PowerPC: Enable MMA for BF16 in llamafile_sgemm
#2 opened Apr 28, 2025 by shalinib-ibm Owner Loading…
Enable MMA for BF16 data type on ppc64le
#1 opened Apr 4, 2025 by shalinib-ibm Owner Loading…
ProTip! Adding no:label will show everything without a label.