vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

llama : multi-threaded quantization (#1075)

Browse source 
* Multi-threading quantization.

Not much gain for simple quantizations, bit it will be important
for quantizations that require more CPU cycles.

* Multi-threading for quantize-stats

It now does the job in ~14 seconds on my Mac for
Q4_0, Q4_1 and Q4_2. Single-threaded it was taking
more than 2 minutes after adding the more elaborate
version of Q4_2.

* Reviewer comments

* Avoiding compiler confusion

After changing chunk_size to const int as suggested by
@ggerganov, clang and GCC starting to warn me that I don't
need to capture it in the lambda. So, I removed it from the
capture list. But that makes the MSVC build fail. So,
making it a constexpr to make every compiler happy.

* Still fighting with lambda captures in MSVC

---------

Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
This commit is contained in:
Kawrakow 2023-04-20 19:42:27 +02:00 • committed by GitHub
parent e0305ead3a
commit 38de86a711
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
6 changed files with 183 additions and 61 deletions
Download patch file Download diff file Expand all files Collapse all files

2
ggml.h
View file

@ -813,6 +813,8 @@ size_t ggml_quantize_q4_1(const float * src, void * dst, int n, int k, int64_t *
size_t ggml_quantize_q4_2(const float * src, void * dst, int n, int k, int64_t * hist);
size_t ggml_quantize_q4_3(const float * src, void * dst, int n, int k, int64_t * hist);
size_t ggml_quantize_chunk(enum ggml_type type, const float * src, void * dst, int start, int n, int64_t * hist);
//
// system info
//