summary refs log tree commit diff stats
path: root/results/classifier/gemma3:27b/instruction/1594069
diff options
context:
space:
mode:
authorChristian Krinitsin <mail@krinitsin.com>2025-07-06 16:43:19 +0000
committerChristian Krinitsin <mail@krinitsin.com>2025-07-06 16:43:19 +0000
commit238ec2b7cc1557d6f34c33cc482e4d0cd3e266dd (patch)
treecd8a1b75ba7b3543eb7fe6857f408e7be4d9fd0b /results/classifier/gemma3:27b/instruction/1594069
parent96049c939b1916d80532630d63c14e04d5244f1d (diff)
downloademulator-bug-study-238ec2b7cc1557d6f34c33cc482e4d0cd3e266dd.tar.gz
emulator-bug-study-238ec2b7cc1557d6f34c33cc482e4d0cd3e266dd.zip
add results
Diffstat (limited to 'results/classifier/gemma3:27b/instruction/1594069')
-rw-r--r--results/classifier/gemma3:27b/instruction/159406911
1 files changed, 11 insertions, 0 deletions
diff --git a/results/classifier/gemma3:27b/instruction/1594069 b/results/classifier/gemma3:27b/instruction/1594069
new file mode 100644
index 00000000..2d47150c
--- /dev/null
+++ b/results/classifier/gemma3:27b/instruction/1594069
@@ -0,0 +1,11 @@
+
+
+
+SIMD instructions translated to scalar host instructions
+
+SIMD instructions inside the guest (NEON, MMX, SSE, SSE2, AVX) are translated to scalar instructions on the host instead of SIMD instructions.  It appears that there have been a few efforts to rectify this [1], and even a submitted patch series, but all discussion has effectively died out [2].
+
+I would like to see better SIMD performance on qemu, especially as non-x86 architectures are becoming widely used (e.g. ARM).
+
+[1] http://dl.acm.org/citation.cfm?id=2757098&dl=ACM&coll=DL&CFID=633095244&CFTOKEN=12352103
+[2] https://lists.nongnu.org/archive/html/qemu-devel/2014-10/msg01720.html
\ No newline at end of file