OV - Compare Loops

Loops

▶mmq.cpp: 1597 - 173.14 %

ASM Loop ID	GFLOP/s	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	GFLOP/s	ASM Loop ID	GFLOP/s	ASM Loop ID	GFLOP/s	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	GFLOP/s	ASM Loop ID	GFLOP/s
Run orig_default		Run gcc_default							Run aocc_default		Run icx_10		Run gcc_4							Run aocc_6
Loop Source Regions		Loop Source Regions	/scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1597-1597 /scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/simd-mappings.h: 130-130 /usr/lib/gcc/x86_64-linux-gnu/13/include/avx512vnniintrin.h: 41-41 /usr/lib/gcc/x86_64-linux-gnu/13/include/avx512bwintrin.h: 958-958 /usr/lib/gcc/x86_64-linux-gnu/13/include/avx512fintrin.h: 261-261 /usr/lib/gcc/x86_64-linux-gnu/13/include/avx512fintrin.h: 954-954 /usr/lib/gcc/x86_64-linux-gnu/13/include/avx512fintrin.h: 4181-4181 /usr/lib/gcc/x86_64-linux-gnu/13/include/avx512fintrin.h: 12821-12821 /usr/lib/gcc/x86_64-linux-gnu/13/include/avx512fintrin.h: 13377-13377 /usr/lib/gcc/x86_64-linux-gnu/13/include/avx512fintrin.h: 14157-14157 /usr/lib/gcc/x86_64-linux-gnu/13/include/avx512fintrin.h: 14630-14630						Loop Source Regions		Loop Source Regions		Loop Source Regions	/scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1597-1597 /scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/simd-mappings.h: 130-130 /usr/lib/gcc/x86_64-linux-gnu/13/include/avx512vnniintrin.h: 41-41 /usr/lib/gcc/x86_64-linux-gnu/13/include/avx512bwintrin.h: 958-958 /usr/lib/gcc/x86_64-linux-gnu/13/include/avx512fintrin.h: 261-261 /usr/lib/gcc/x86_64-linux-gnu/13/include/avx512fintrin.h: 954-954 /usr/lib/gcc/x86_64-linux-gnu/13/include/avx512fintrin.h: 4181-4181 /usr/lib/gcc/x86_64-linux-gnu/13/include/avx512fintrin.h: 12821-12821 /usr/lib/gcc/x86_64-linux-gnu/13/include/avx512fintrin.h: 13377-13377 /usr/lib/gcc/x86_64-linux-gnu/13/include/avx512fintrin.h: 14157-14157 /usr/lib/gcc/x86_64-linux-gnu/13/include/avx512fintrin.h: 14630-14630						Loop Source Regions
		405	24.65	24.14	86.35	87.67	85.7	89.25					400	24.73	24.14	86.79	87.67	85.7	89.52

No Optimizer analysis found for any assembly loop. More loops can be analyzed using option --optimizer-loop-count.		Sum on 1 analyzed binary loop (libggml-cpu.so - 405)							No Optimizer analysis found for any assembly loop. More loops can be analyzed using option --optimizer-loop-count.		No Optimizer analysis found for any assembly loop. More loops can be analyzed using option --optimizer-loop-count.		Sum on 1 analyzed binary loop (libggml-cpu.so - 400)							No Optimizer analysis found for any assembly loop. More loops can be analyzed using option --optimizer-loop-count.
Analysis	Count	Analysis						Count	Analysis	Count	Analysis	Count	Analysis						Count	Analysis	Count
		Data Access Issues											Data Access Issues
		Presence of constant non-unit stride data access						1					Presence of constant non-unit stride data access						1
		Presence of indirect access						1					Presence of indirect access						1
		Presence of special instructions executing on a single port						1					Presence of special instructions executing on a single port						1
		Vectorization Roadblocks											Vectorization Roadblocks
		Presence of constant non-unit stride data access						1					Presence of constant non-unit stride data access						1
		Presence of indirect access						1					Presence of indirect access						1
		Inefficient Vectorization											Inefficient Vectorization
		Presence of special instructions executing on a single port						1					Presence of special instructions executing on a single port						1

▶mmq.cpp: 1570 - 170.17 %

ASM Loop ID	GFLOP/s	ASM Loop ID	GFLOP/s	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	GFLOP/s	ASM Loop ID	GFLOP/s	ASM Loop ID	GFLOP/s	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	GFLOP/s
Run orig_default		Run gcc_default		Run aocc_default							Run icx_10		Run gcc_4		Run aocc_6
Loop Source Regions		Loop Source Regions		Loop Source Regions	/scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1570-1575 /scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1584-1584 /scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1590-1597						Loop Source Regions		Loop Source Regions		Loop Source Regions	/scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1570-1575 /scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1584-1584 /scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1590-1597
				563	24.50	23.97	84.92	0	0	89.54					511	24.58	23.92	85.25	0	0	90.49

No Optimizer analysis found for any assembly loop. More loops can be analyzed using option --optimizer-loop-count.		No Optimizer analysis found for any assembly loop. More loops can be analyzed using option --optimizer-loop-count.		Sum on 1 analyzed binary loop (libggml-cpu.so - 563)							No Optimizer analysis found for any assembly loop. More loops can be analyzed using option --optimizer-loop-count.		No Optimizer analysis found for any assembly loop. More loops can be analyzed using option --optimizer-loop-count.		Sum on 1 analyzed binary loop (libggml-cpu.so - 511)
Analysis	Count	Analysis	Count	Analysis						Count	Analysis	Count	Analysis	Count	Analysis						Count

▶mmq.cpp: 1573 - 169.87 %

ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	GFLOP/s	ASM Loop ID	GFLOP/s	ASM Loop ID	GFLOP/s	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	GFLOP/s	ASM Loop ID	GFLOP/s	ASM Loop ID	GFLOP/s
Run orig_default							Run gcc_default		Run aocc_default		Run icx_10							Run gcc_4		Run aocc_6
Loop Source Regions	/scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1573-1575 /scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1584-1584 /scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1590-1597						Loop Source Regions		Loop Source Regions		Loop Source Regions	/scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1573-1575 /scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1584-1584 /scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1590-1597						Loop Source Regions		Loop Source Regions
452	24.60	24.02	84.98	0	0	89.29					322	24.63	24.01	84.90	0	0	89.34

Sum on 1 analyzed binary loop (libggml-cpu.so - 452)							No Optimizer analysis found for any assembly loop. More loops can be analyzed using option --optimizer-loop-count.		No Optimizer analysis found for any assembly loop. More loops can be analyzed using option --optimizer-loop-count.		Sum on 1 analyzed binary loop (libggml-cpu.so - 322)							No Optimizer analysis found for any assembly loop. More loops can be analyzed using option --optimizer-loop-count.		No Optimizer analysis found for any assembly loop. More loops can be analyzed using option --optimizer-loop-count.
Analysis						Count	Analysis	Count	Analysis	Count	Analysis						Count	Analysis	Count	Analysis	Count

▶<unknown>: 0 - 6.31 %

ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	GFLOP/s	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	GFLOP/s	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	GFLOP/s	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	GFLOP/s	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	GFLOP/s	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	GFLOP/s
Run orig_default							Run gcc_default							Run aocc_default							Run icx_10							Run gcc_4							Run aocc_6
Loop Source Regions							Loop Source Regions							Loop Source Regions							Loop Source Regions							Loop Source Regions							Loop Source Regions
1900	0.00	0.00	0.00	0	0	0	1407	0.01	0.00	0.00	0	0	0	1138	0.00	0.00	0.00	0	0	0	2016	0.00	0.00	0.00	0	0	0	1610	0.00	0.00	0.00	0	0	0	2360	0.00	0.00	0.00	0	0	0
2328	0.01	0.00	0.00	0	0	0	3040	0.00	0.00	0.00	0	0	0	2873	0.01	0.00	0.00	0	0	0	2316	0.01	0.00	0.00	0	0	0	3089	0.01	0.00	0.00	0	0	0	2254	0.00	0.00	0.00	0	0	0
2321	0.00	0.00	0.00	0	0	0	3036	0.01	0.00	0.00	0	0	0	2870	0.01	0.00	0.00	0	0	0	996	0.00	0.00	0.00	0	0	0	3102	0.00	0.00	0.00	0	0	0	2605	0.01	0.00	0.00	0	0	0
2552	0.01	0.00	0.00	0	0	0	2738	0.08	0.00	0.01	0	0	0	1097	0.00	0.00	0.00	0	0	0	285	0.00	0.00	0.00	0	0	0	2802	0.08	0.00	0.01	0	0	0	2268	0.01	0.00	0.00	0	0	0
1939	0.00	0.00	0.00	0	0	NA	1409	0.00	0.00	0.00	0	0	0	1256	0.00	0.00	0.00	0	0	0	1062	0.01	0.00	0.00	0	0	0	1802	0.01	0.00	0.00	0	0	0	2316	0.00	0.00	0.00	0	0	0
1021	0.00	0.00	0.00	0	0	0	3463	0.01	0.00	0.00	0	0	0	978	0.00	0.00	0.00	0	0	0	1702	0.01	0.00	0.00	0	0	0	3558	0.01	0.00	0.00	0	0	0	1101	0.00	0.00	0.00	0	0	0
2334	0.02	0.00	0.00	0	0	0	551	0.00	0.00	0.00	0	0	0	2749	0.01	0.00	0.00	0	0	0	2111	0.00	0.00	0.00	0	0	0	3029	0.01	0.00	0.00	0	0	0	2503	0.01	0.00	0.00	0	0	0
2146	0.00	0.00	0.00	0	0	0	3208	0.01	0.00	0.00	0	0	0	2486	0.00	0.00	0.00	0	0	0	2114	0.02	0.00	0.00	0	0	0	3257	0.02	0.00	0.00	0	0	0	2489	0.01	0.00	0.00	0	0	0
2448	0.00	0.00	0.00	0	0	0	3178	0.01	0.00	0.00	0	0	0	2733	0.01	0.00	0.00	0	0	0	1061	0.00	0.00	0.00	0	0	0	2734	0.00	0.00	0.00	0	0	0	926	0.00	0.00	0.00	0	0	0
2167	0.01	0.00	0.00	0	0	0	1369	0.00	0.00	0.00	0	0	0	2461	0.00	0.00	0.00	0	0	0	926	0.01	0.00	0.00	0	0	0	3098	0.00	0.00	0.00	0	0	0	1952	0.00	0.00	0.00	0	0	0
2437	0.00	0.00	0.00	0	0	0	2964	0.01	0.00	0.00	0	0	0	1033	0.00	0.00	0.00	0	0	0	1953	0.00	0.00	0.00	0	0	0	2144	0.00	0.00	0.00	0	0	0	1970	0.01	0.00	0.00	0	0	0
2176	0.01	0.00	0.00	0	0	0	2675	0.00	0.00	0.00	0	0	0	2871	0.00	0.00	0.00	0	0	0	1962	0.00	0.00	0.00	0	0	0	1468	0.01	0.00	0.00	0	0	0	1136	0.00	0.00	0.00	0	0	0
1195	0.00	0.00	0.00	0	0	0	2938	0.02	0.00	0.00	0	0	0	2478	0.00	0.00	0.00	0	0	0	1974	0.01	0.00	0.00	0	0	0	3232	0.00	0.00	0.00	0	0	0	896	0.00	0.00	0.00	0	0	0
2438	0.01	0.00	0.00	0	0	0	2936	0.01	0.00	0.00	0	0	0	2492	0.01	0.00	0.00	0	0	2.7	1955	0.00	0.00	0.00	0	0	0	2803	0.00	0.00	0.00	0	0	0	2499	0.02	0.00	0.00	0	0	0
2322	0.01	0.00	0.00	0	0	0	1549	0.00	0.00	0.00	0	0	0	2891	0.00	0.00	0.00	0	0	0	1981	0.00	0.00	0.00	0	0	0	1844	0.00	0.00	0.00	0	0	0	2262	0.00	0.00	0.00	0	0	0
2435	0.01	0.00	0.00	0	0	0	3149	0.01	0.00	0.00	0	0	0	2215	0.01	0.00	0.00	0	0	0	1719	0.01	0.00	0.00	0	0	0	3003	0.02	0.00	0.00	0	0	0	1054	0.00	0.00	0.00	0	0	0
1901	0.01	0.00	0.00	0	0	0	3153	0.01	0.00	0.00	0	0	0	1140	0.01	0.00	0.00	0	0	0	1715	0.00	0.00	0.00	0	0	0	3448	0.00	0.00	0.00	0	0	0	972	0.00	0.00	0.00	0	0	0
2155	0.00	0.00	0.00	0	0	0	2939	0.01	0.00	0.00	0	0	0	2896	0.00	0.00	0.00	0	0	0	1027	0.00	0.00	0.00	0	0	0	3289	0.02	0.00	0.00	0	0	0	2602	0.00	0.00	0.00	0	0	0
2174	0.00	0.00	0.00	0	0	0	3357	0.00	0.00	0.00	0	0	0	2744	0.01	0.00	0.00	0	0	0	2201	0.01	0.00	0.00	0	0	0	3004	0.00	0.00	0.00	0	0	0	445	0.00	0.00	0.00	0	0	0
484	0.01	0.00	0.00	0	0	0	2674	0.00	0.00	0.00	0	0	0	2275	0.06	0.00	0.01	0	0	0	2110	0.00	0.00	0.00	0	0	0	3095	0.00	0.00	0.00	0	0	0	2236	0.00	0.00	0.00	0	0	0
504	0.00	0.00	0.00	0	0	0	1422	0.00	0.00	0.00	0	0	0	530	0.01	0.00	0.00	0	0	0	2204	0.01	0.00	0.00	0	0	0	3228	0.01	0.00	0.00	0	0	0	2042	0.05	0.00	0.00	0	0	0
483	0.01	0.00	0.00	0	0	0	727	0.00	0.00	0.00	0	0	0	459	0.01	0.00	0.00	0	0	0	1983	0.02	0.00	0.00	0	0	0	3006	0.01	0.00	0.00	0	0	0	390	0.01	0.00	0.00	0	0	0
1868	0.05	0.00	0.00	0	0	0	3228	0.15	0.00	0.01	0	0	0	877	0.00	0.00	0.00	0	0	0	2116	0.01	0.00	0.00	0	0	0	752	0.00	0.00	0.00	0	0	0	391	0.00	0.00	0.00	0	0	0
2004	0.01	0.00	0.00	0	0	0	31	0.00	0.00	0.00	0	0	0	325	0.00	0.00	0.00	0	0	56.8	1647	0.07	0.00	0.01	0	0	0	2377	0.00	0.00	0.00	0	0	0	462	0.00	0.00	0.00	0	0	0
2003	0.01	0.00	0.00	0	0	131.92	471	0.01	0.00	0.00	0	0	150.61	991	0.00	0.00	0.00	0	0	0	443	0.00	0.00	0.00	0	0	0	3381	0.12	0.00	0.01	0	0	0	763	0.00	0.00	0.00	0	0	0
452	0.00	0.00	0.00	0	0	0	706	0.00	0.00	0.00	0	0	0	734	0.00	0.00	0.00	0	0	604.97	413	0.01	0.00	0.00	0	0	0	651	0.00	0.00	0.00	0	0	0	333	0.01	0.00	0.00	0	0	182.13
804	0.00	0.00	0.00	0	0	0	510	0.03	0.01	0.03	0	0	1507.18	416	0.14	0.10	0.37	0	0	0	463	0.01	0.00	0.00	0	0	0	677	0.00	0.00	0.00	0	0	0	816	0.00	0.00	0.00	0	0	0
700	0.01	0.00	0.00	0	0	0	636	0.09	0.00	0.01	0	0	520.25	1790	0.15	0.00	0.01	0	0	410.49	1738	0.03	0.00	0.00	0	0	182.88	441	0.01	0.00	0.00	0	0	0	83	0.00	0.00	0.00	0	0	NA
365	0.01	0.00	0.00	0	0	75.34	1348	0.01	0.00	0.00	0	0	0	1264	0.24	0.08	0.27	0	0	1885.36	633	0.00	0.00	0.00	0	0	0	218	0.01	0.00	0.00	0	0	0	1747	0.00	0.00	0.00	0	0	0
161	0.00	0.00	0.00	0	0	0	312	0.02	0.00	0.01	0	0	0	2137	0.02	0.00	0.01	0	0	600.88	734	0.00	0.00	0.00	0	0	0	221	0.00	0.00	0.00	0	0	0	387	0.00	0.00	0.00	0	0	NA
639	0.17	0.10	0.37	0	0	0	313	0.00	0.00	0.00	0	0	0	107	0.00	0.00	0.00	0	0	NA	316	0.01	0.00	0.00	0	0	188.54	66	0.00	0.00	0.00	0	0	NA	2029	0.02	0.01	0.02	0	0	0.1
122	0.03	0.01	0.03	0	0	19.23	52	0.04	0.01	0.05	0	0	10.36	727	0.00	0.00	0.00	0	0	0	19	0.01	0.00	0.00	0	0	0	535	0.01	0.00	0.00	0	0	904.33	1279	0.25	0.08	0.29	0	0	1768.76
813	0.20	0.00	0.01	0	0	206.73	459	0.12	0.07	0.25	0	0	0	739	0.01	0.00	0.00	0	0	286.73	696	0.17	0.00	0.02	0	0	190.48	615	0.00	0.00	0.00	0	0	0	689	0.01	0.00	0.00	0	0	4060.6
2176	0.01	0.00	0.00	0	0	301.62	1684	0.00	0.00	0.00	0	0	NA	737	0.01	0.00	0.00	0	0	245.94	1089	0.07	0.03	0.12	0	0	1429.5	530	0.01	0.00	0.00	0	0	5235.02	358	0.14	0.11	0.38	0	0	0
116	0.01	0.00	0.00	0	0	0	70	0.00	0.00	0.00	0	0	NA	108	0.04	0.01	0.03	0	0	43.77	1096	0.16	0.00	0.01	0	0	8015.27	532	0.01	0.00	0.00	0	0	552.3	2047	0.14	0.07	0.24	0	0	880.52
451	0.01	0.00	0.00	0	0	385.3	1343	0.00	0.00	0.00	0	0	0	741	0.01	0.00	0.00	0	0	4142.9	1518	0.17	0.00	0.01	0	0	410.51	1307	0.07	0.00	0.00	0	0	1259.28	687	0.01	0.00	0.00	0	0	361.27
1573	0.01	0.00	0.00	0	0	7691.55	514	0.00	0.00	0.00	0	0	0	410	0.00	0.00	0.00	0	0	0	1512	0.00	0.00	0.00	0	0	301.43	537	0.01	0.00	0.00	0	0	0	1647	0.03	0.00	0.00	0	0	2609.14
2191	0.00	0.00	0.00	0	0	496.41	67	0.01	0.00	0.00	0	0	0	2473	0.00	0.00	0.00	0	0	151.44	321	0.01	0.00	0.00	0	0	276.44	536	0.00	0.00	0.00	0	0	8433.28	683	0.00	0.00	0.00	0	0	NA
6	0.02	0.00	0.01	0	0	0	980	0.02	0.00	0.02	0	0	220.27	848	0.09	0.00	0.01	0	0	432.01	2080	0.00	0.00	0.00	0	0	NA	63	0.01	0.00	0.00	0	0	0	2051	0.00	0.00	0.00	0	0	NA
1812	0.02	0.00	0.02	0	0	303.17	985	0.03	0.01	0.03	0	0	669.16	5	0.01	0.00	0.01	0	0	0	52	0.00	0.00	0.00	0	0	0	1627	0.16	0.09	0.31	0	0	1364.7	352	0.01	0.00	0.00	0	0	0
547	0.00	0.00	0.00	0	0	0	1593	0.01	0.00	0.00	0	0	204.39	2469	0.14	0.07	0.24	0	0	884.41	541	0.14	0.11	0.38	0	0	0	729	0.15	0.00	0.01	0	0	229.44	681	0.00	0.00	0.00	0	0	0
1837	0.03	0.01	0.03	0	0	305.41	1599	0.15	0.08	0.28	0	0	926.47	2145	0.03	0.01	0.04	0	0	415.06	6	0.01	0.00	0.01	0	0	0	1367	0.00	0.00	0.00	0	0	0	2762	0.24	0.01	0.02	0	0	1218.17
405	0.00	0.00	0.00	0	0	0	1283	0.00	0.00	0.00	0	0	11749.07	2451	0.03	0.01	0.02	0	0	10.72	282	0.00	0.00	0.00	0	0	0	48	0.04	0.01	0.05	0	0	13.18	1289	0.13	0.00	0.01	0	0	7659.04
553	0.02	0.00	0.00	0	0	287.87	513	0.01	0.00	0.00	0	0	3916.64	2442	0.01	0.00	0.00	0	0	423.96	1956	0.02	0.01	0.02	0	0	2.51	1306	0.01	0.00	0.00	0	0	5955.1	2024	0.00	0.00	0.00	0	0	285.65
555	0.00	0.00	0.00	0	0	0	952	0.14	0.00	0.01	0	0	7814.8	1791	0.01	0.00	0.00	0	0	5480.17	1952	0.01	0.00	0.00	0	0	337.63	3	0.02	0.00	0.01	0	0	0	84	0.03	0.01	0.02	0	0	23.24
557	0.01	0.00	0.01	0	0	93.6	1284	0.20	0.00	0.02	0	0	364.06	560	0.00	0.00	0.00	0	0	778.49	1964	0.21	0.11	0.39	0	0	1404.31	1004	0.02	0.00	0.02	0	0	188.18	1836	0.02	0.00	0.02	0	0	241.36
2180	0.02	0.01	0.02	0	0	1.94	1596	0.03	0.01	0.02	0	0	0	736	0.01	0.00	0.00	0	0	86.5	57	0.00	0.00	0.00	0	0	50.27	655	0.05	0.00	0.01	0	0	614.8	76	0.00	0.00	0.00	0	0	0
2189	0.19	0.11	0.39	0	0	1293.28	3	0.01	0.00	0.01	0	0	0	738	0.00	0.00	0.00	0	0	454.13	281	0.02	0.00	0.01	0	0	4.43	589	0.01	0.00	0.00	0	0	0	1646	0.01	0.00	0.00	0	0	5025.11
548	0.02	0.00	0.00	0	0	2459.7	948	0.15	0.08	0.27	0	0	2127.28	1268	0.13	0.00	0.01	0	0	9120.76	1725	0.04	0.01	0.03	0	0	298.84	1624	0.02	0.01	0.02	0	0	1.38	814	0.08	0.00	0.01	0	0	334.6
1578	0.18	0.00	0.01	0	0	381.47	509	0.01	0.00	0.01	0	0	374.37	946	0.16	0.00	0.02	0	0	185.65	58	0.04	0.01	0.03	0	0	7.82	974	0.13	0.00	0.01	0	0	8366.06	685	0.00	0.00	0.00	0	0	474.25
1120	0.09	0.03	0.12	0	0	1553.25	1601	0.01	0.00	0.00	0	0	161.92	2632	0.00	0.00	0.00	0	0	302.48	1967	0.01	0.00	0.00	0	0	1017.95	972	0.08	0.04	0.15	0	0	1555.78	1840	0.04	0.01	0.05	0	0	332.76
1127	0.13	0.00	0.01	0	0	7959.93	710	0.13	0.00	0.01	0	0	199.93	409	0.01	0.00	0.01	0	0	4.8	562	0.13	0.00	0.01	0	0	307.44	1629	0.00	0.00	0.00	0	0	542.87	684	0.01	0.00	0.00	0	0	227.6
404	0.01	0.00	0.01	0	0	0															1516	0.01	0.00	0.00	0	0	5145.77	1009	0.02	0.01	0.03	0	0	592.78	1452	0.00	0.00	0.00	0	0	0
699	0.09	0.00	0.01	0	0	379.85															448	0.02	0.00	0.00	0	0	239.89	476	0.12	0.07	0.26	0	0	0	686	0.01	0.00	0.00	0	0	433.54
																					1713	0.03	0.00	0.02	0	0	299.62								5	0.01	0.00	0.00	0	0	0
																					447	0.02	0.00	0.01	0	0	80.11								351	0.01	0.00	0.01	0	0	11.04
																					446	0.01	0.00	0.00	0	0	3399.33								906	0.18	0.00	0.02	0	0	168.97
																																			508	0.01	0.00	0.00	0	0	520.31

No Optimizer analysis found for any assembly loop. More loops can be analyzed using option --optimizer-loop-count.							No Optimizer analysis found for any assembly loop. More loops can be analyzed using option --optimizer-loop-count.							No Optimizer analysis found for any assembly loop. More loops can be analyzed using option --optimizer-loop-count.							No Optimizer analysis found for any assembly loop. More loops can be analyzed using option --optimizer-loop-count.							No Optimizer analysis found for any assembly loop. More loops can be analyzed using option --optimizer-loop-count.							No Optimizer analysis found for any assembly loop. More loops can be analyzed using option --optimizer-loop-count.
Analysis						Count	Analysis						Count	Analysis						Count	Analysis						Count	Analysis						Count	Analysis						Count

▶quants.c: 298 - 0.14 %

ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	GFLOP/s	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	GFLOP/s	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	GFLOP/s	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	GFLOP/s	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	GFLOP/s	ASM Loop ID	GFLOP/s
Run orig_default							Run gcc_default							Run aocc_default							Run icx_10							Run gcc_4							Run aocc_6
Loop Source Regions	/scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/arch/x86/quants.c: 298-347 /scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/arch/x86/quants.c: 353-355 /scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/./ggml-impl.h: 346-346 /scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/./ggml-impl.h: 355-355 /scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/./ggml-impl.h: 389-404						Loop Source Regions	/usr/lib/gcc/x86_64-linux-gnu/13/include/pmmintrin.h: 71-71 /usr/lib/gcc/x86_64-linux-gnu/13/include/avxintrin.h: 186-186 /usr/lib/gcc/x86_64-linux-gnu/13/include/avxintrin.h: 296-296 /usr/lib/gcc/x86_64-linux-gnu/13/include/avxintrin.h: 320-320 /usr/lib/gcc/x86_64-linux-gnu/13/include/avxintrin.h: 475-475 /usr/lib/gcc/x86_64-linux-gnu/13/include/avxintrin.h: 526-526 /usr/lib/gcc/x86_64-linux-gnu/13/include/avxintrin.h: 905-905 /usr/lib/gcc/x86_64-linux-gnu/13/include/avxintrin.h: 935-935 /usr/lib/gcc/x86_64-linux-gnu/13/include/avxintrin.h: 1066-1066 /usr/lib/gcc/x86_64-linux-gnu/13/include/avxintrin.h: 1329-1329 /usr/lib/gcc/x86_64-linux-gnu/13/include/avxintrin.h: 1472-1472 /usr/lib/gcc/x86_64-linux-gnu/13/include/avx2intrin.h: 79-79 /usr/lib/gcc/x86_64-linux-gnu/13/include/avx2intrin.h: 86-86 /usr/lib/gcc/x86_64-linux-gnu/13/include/avx2intrin.h: 1046-1046 /scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/arch/x86/quants.c: 298-298 /scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/arch/x86/quants.c: 304-304 /scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/arch/x86/quants.c: 319-321 /scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/./ggml-impl.h: 354-354 /scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/./ggml-impl.h: 389-404 /usr/lib/gcc/x86_64-linux-gnu/13/include/xmmintrin.h: 184-184 /usr/lib/gcc/x86_64-linux-gnu/13/include/xmmintrin.h: 240-240 /usr/lib/gcc/x86_64-linux-gnu/13/include/xmmintrin.h: 793-793						Loop Source Regions	/scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/arch/x86/quants.c: 298-347 /scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/arch/x86/quants.c: 353-355 /scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/./ggml-impl.h: 346-346 /scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/./ggml-impl.h: 355-355 /scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/./ggml-impl.h: 389-404						Loop Source Regions	/scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/arch/x86/quants.c: 298-347 /scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/arch/x86/quants.c: 353-355 /scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/./ggml-impl.h: 346-346 /scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/./ggml-impl.h: 355-355 /scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/./ggml-impl.h: 389-404						Loop Source Regions	/usr/lib/gcc/x86_64-linux-gnu/13/include/pmmintrin.h: 71-71 /usr/lib/gcc/x86_64-linux-gnu/13/include/avxintrin.h: 186-186 /usr/lib/gcc/x86_64-linux-gnu/13/include/avxintrin.h: 296-296 /usr/lib/gcc/x86_64-linux-gnu/13/include/avxintrin.h: 320-320 /usr/lib/gcc/x86_64-linux-gnu/13/include/avxintrin.h: 475-475 /usr/lib/gcc/x86_64-linux-gnu/13/include/avxintrin.h: 526-526 /usr/lib/gcc/x86_64-linux-gnu/13/include/avxintrin.h: 905-905 /usr/lib/gcc/x86_64-linux-gnu/13/include/avxintrin.h: 935-935 /usr/lib/gcc/x86_64-linux-gnu/13/include/avxintrin.h: 1066-1066 /usr/lib/gcc/x86_64-linux-gnu/13/include/avxintrin.h: 1329-1329 /usr/lib/gcc/x86_64-linux-gnu/13/include/avx2intrin.h: 79-79 /usr/lib/gcc/x86_64-linux-gnu/13/include/avx2intrin.h: 86-86 /usr/lib/gcc/x86_64-linux-gnu/13/include/avx2intrin.h: 1046-1046 /scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/arch/x86/quants.c: 298-298 /scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/arch/x86/quants.c: 304-304 /scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/arch/x86/quants.c: 319-321 /scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/./ggml-impl.h: 354-354 /scratch/users/amazouz/QAAS/service/Llama.cpp/sdp772511/175-793-6543/llama.cpp/build/llama.cpp/ggml/src/./ggml-impl.h: 389-404 /usr/lib/gcc/x86_64-linux-gnu/13/include/xmmintrin.h: 184-184 /usr/lib/gcc/x86_64-linux-gnu/13/include/xmmintrin.h: 240-240 /usr/lib/gcc/x86_64-linux-gnu/13/include/xmmintrin.h: 793-793						Loop Source Regions
2925	0.28	0.01	0.03	60.7	29.66	1147.06	2122	0.36	0.01	0.03	60	28.75	1070.9	3339	0.29	0.01	0.03	58.33	28.75	1046.81	2663	0.35	0.01	0.03	60.7	29.66	1043.67	2154	0.31	0.01	0.03	59.65	29.28	1046.73

Sum on 1 analyzed binary loop (libggml-cpu.so - 2925)							Sum on 1 analyzed binary loop (libggml-cpu.so - 2122)							Sum on 1 analyzed binary loop (libggml-cpu.so - 3339)							Sum on 1 analyzed binary loop (libggml-cpu.so - 2663)							Sum on 1 analyzed binary loop (libggml-cpu.so - 2154)							No Optimizer analysis found for any assembly loop. More loops can be analyzed using option --optimizer-loop-count.
Analysis						Count	Analysis						Count	Analysis						Count	Analysis						Count	Analysis						Count	Analysis	Count
Loop Computation Issues							Loop Computation Issues							Loop Computation Issues							Loop Computation Issues							Loop Computation Issues
Presence of expensive FP instructions						1	Presence of expensive FP instructions						1	Presence of expensive FP instructions						1	Presence of expensive FP instructions						1	Presence of expensive FP instructions						1
Less than 10% of the FP ADD/SUB/MUL arithmetic operations are performed using FMA						0	Less than 10% of the FP ADD/SUB/MUL arithmetic operations are performed using FMA						0	Less than 10% of the FP ADD/SUB/MUL arithmetic operations are performed using FMA						1	Less than 10% of the FP ADD/SUB/MUL arithmetic operations are performed using FMA						0	Less than 10% of the FP ADD/SUB/MUL arithmetic operations are performed using FMA						0
Control Flow Issues							Control Flow Issues							Control Flow Issues							Control Flow Issues							Control Flow Issues
Presence of 2 to 4 paths						1	Presence of 2 to 4 paths						0	Presence of 2 to 4 paths						0	Presence of 2 to 4 paths						1	Presence of 2 to 4 paths						0
Presence of more than 4 paths						0	Presence of more than 4 paths						1	Presence of more than 4 paths						1	Presence of more than 4 paths						0	Presence of more than 4 paths						1
Data Access Issues							Data Access Issues							Data Access Issues							Data Access Issues							Data Access Issues
More than 10% of the vector loads instructions are unaligned						1	More than 10% of the vector loads instructions are unaligned						1	More than 10% of the vector loads instructions are unaligned						1	More than 10% of the vector loads instructions are unaligned						1	More than 10% of the vector loads instructions are unaligned						1
Presence of special instructions executing on a single port						1	Presence of special instructions executing on a single port						1	Presence of special instructions executing on a single port						1	Presence of special instructions executing on a single port						1	Presence of special instructions executing on a single port						1
Vectorization Roadblocks							Vectorization Roadblocks							Vectorization Roadblocks							Vectorization Roadblocks							Vectorization Roadblocks
Presence of 2 to 4 paths						1	Presence of 2 to 4 paths						0	Presence of 2 to 4 paths						0	Presence of 2 to 4 paths						1	Presence of 2 to 4 paths						0
Presence of more than 4 paths						0	Presence of more than 4 paths						1	Presence of more than 4 paths						1	Presence of more than 4 paths						0	Presence of more than 4 paths						1
Inefficient Vectorization							Inefficient Vectorization							Inefficient Vectorization							Inefficient Vectorization							Inefficient Vectorization
Presence of special instructions executing on a single port						1	Presence of special instructions executing on a single port						1	Presence of special instructions executing on a single port						1	Presence of special instructions executing on a single port						1	Presence of special instructions executing on a single port						1

Report Configuration

Loops

▶mmq.cpp: 1597 - 173.14 %

▶mmq.cpp: 1570 - 170.17 %

▶mmq.cpp: 1573 - 169.87 %

▶<unknown>: 0 - 6.31 %

▶quants.c: 298 - 0.14 %