ikawrakow/ik_llama.cpp
1
0
Fork 0
mirror of https://github.com/ikawrakow/ik_llama.cpp.git synced 2026-01-26 17:20:01 +00:00
Code Issues Packages Projects Releases Wiki Activity

Fix ARM_NEON build failure due to q8_2 (#303)

Browse Source 
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
This commit is contained in:
Kawrakow
2025-04-01 13:48:20 +02:00
committed by GitHub
parent 1bc60d6cc9
commit df20261b6a
2 changed files with 4 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
ggml/src/ggml-alloc.c
View File

@@ -174,8 +174,8 @@ static size_t ggml_dyn_tallocr_alloc(struct ggml_dyn_tallocr * alloc, size_t siz
// this should never happen
fprintf(stderr, "%s: not enough space in the buffer to allocate %zu bytes, largest block available %zu bytes\n",
__func__, size, max_avail);
fprintf(stderr, "%s: tensor was %s with %zu elements and %zu bytes\n", __func__, tensor->name,
ggml_nelements(tensor), ggml_nbytes(tensor));
fprintf(stderr, "%s: tensor was %s with %g elements and %zu bytes\n", __func__, tensor->name,
1.*ggml_nelements(tensor), ggml_nbytes(tensor));
GGML_ABORT("not enough space in the buffer");
}
}

4
ggml/src/iqk/iqk_quantize.cpp
View File

@@ -860,9 +860,9 @@ void quantize_row_q8_1_x4_T(const float * x, Block * y, int64_t k) {
}
} else {
if (i < nb4) {
y4[i4].s[ir] = vaddvq_s32(accv);
y4[i4].d[ir+4] = GGML_FP32_TO_BF16(d * vaddvq_s32(accv)).bits;
} else {
y[i].s = vaddvq_s32(accv);
y[i].s = GGML_FP32_TO_BF16(d * vaddvq_s32(accv)).bits;
}
}
}