OV - Compare Loops

Loops

▶sgemm.cpp: 138 - 163.87 %

ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	ASM Loop ID	Vector Length Use (%)
Run orig_default						Run icx_default						Run aocc_9						Run icx_2
Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/llamafile/sgemm.cpp: 138-138 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/llamafile/sgemm.cpp: 818-846 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/llamafile/sgemm.cpp: 1038-1038 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/llamafile/sgemm.cpp: 1044-1044					Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/llamafile/sgemm.cpp: 138-138 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/llamafile/sgemm.cpp: 818-821 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/llamafile/sgemm.cpp: 827-846 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/llamafile/sgemm.cpp: 966-966 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/llamafile/sgemm.cpp: 1038-1038 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/llamafile/sgemm.cpp: 1044-1044					Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/llamafile/sgemm.cpp: 138-138 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/llamafile/sgemm.cpp: 826-846 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/llamafile/sgemm.cpp: 966-966 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/llamafile/sgemm.cpp: 1038-1038 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/llamafile/sgemm.cpp: 1044-1044					Loop Source Regions
2364	8.72	6.41	64.06	0	0	2688	8.60	6.36	63.92	0	0	1814	5.26	3.59	35.89	0	0

Sum on 1 analyzed binary loop (libggml-cpu.so - 2364)						Sum on 1 analyzed binary loop (libggml-cpu.so - 2688)						Sum on 1 analyzed binary loop (libggml-cpu.so - 1814)						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

▶quants.c: 1066 - 78.11 %

ASM Loop ID	Vector Length Use (%)	ASM Loop ID	Vector Length Use (%)	ASM Loop ID	Vector Length Use (%)	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)
Run orig_default		Run icx_default		Run aocc_9		Run icx_2
Loop Source Regions		Loop Source Regions		Loop Source Regions		Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/simd-mappings.h: 130-130 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/arch/x86/quants.c: 1066-1073
						4152	11.09	10.24	78.11	88.89	20.14

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.		Sum on 1 analyzed binary loop (libggml-cpu.so - 4152)
Analysis	Count	Analysis	Count	Analysis	Count	Analysis					Count
						Loop Computation Issues
						Less than 10% of the FP ADD/SUB/MUL arithmetic operations are performed using FMA					1
						Data Access Issues
						Presence of indirect access					1
						Presence of special instructions executing on a single port					1
						Vectorization Roadblocks
						Presence of indirect access					1
						Inefficient Vectorization
						Presence of special instructions executing on a single port					1

▶quants.c: 298 - 1.71 %

ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	ASM Loop ID	Vector Length Use (%)
Run orig_default						Run icx_default						Run aocc_9						Run icx_2
Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/./ggml-impl.h: 346-346 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/./ggml-impl.h: 355-355 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/./ggml-impl.h: 389-404 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/arch/x86/quants.c: 298-347 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/arch/x86/quants.c: 353-355					Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/./ggml-impl.h: 346-346 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/./ggml-impl.h: 355-355 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/./ggml-impl.h: 389-404 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/arch/x86/quants.c: 298-347 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/arch/x86/quants.c: 353-355					Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/./ggml-impl.h: 346-346 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/./ggml-impl.h: 355-355 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/./ggml-impl.h: 389-404 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/arch/x86/quants.c: 298-347 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/arch/x86/quants.c: 353-355					Loop Source Regions
2527	0.12	0.06	0.58	58.66	28.88	2918	0.11	0.06	0.58	59.66	29.26	1951	0.12	0.05	0.54	60.7	29.66

Sum on 1 analyzed binary loop (libggml-cpu.so - 2527)						Sum on 1 analyzed binary loop (libggml-cpu.so - 2918)						Sum on 1 analyzed binary loop (libggml-cpu.so - 1951)						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
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
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
Presence of 2 to 4 paths					1	Presence of 2 to 4 paths					1	Presence of 2 to 4 paths					1
Data Access Issues						Data Access Issues						Data Access Issues
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
Presence of 2 to 4 paths					1	Presence of 2 to 4 paths					1	Presence of 2 to 4 paths					1
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

▶<unknown>: 0 - 1.50 %

ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)
Run orig_default						Run icx_default						Run aocc_9						Run icx_2
Loop Source Regions						Loop Source Regions						Loop Source Regions						Loop Source Regions
2738	0.01	0.00	0.00	0	0	2395	0.00	0.00	0.00	0	0	2157	0.00	0.00	0.00	0	0	4502	0.01	0.00	0.00	0	0
2475	0.01	0.00	0.00	0	0	2504	0.00	0.00	0.00	0	0	2349	0.02	0.00	0.00	0	0	1896	0.01	0.00	0.00	0	0
2640	0.01	0.00	0.00	0	0	1947	0.03	0.00	0.00	0	0	2262	0.00	0.00	0.00	0	0	4206	0.01	0.00	0.00	0	0
2458	0.01	0.00	0.00	0	0	192	0.00	0.00	0.00	0	0	1058	0.00	0.00	0.00	0	0	5002	0.01	0.00	0.00	0	0
2842	0.01	0.00	0.00	0	0	1948	0.01	0.00	0.01	0	0	2448	0.01	0.00	0.00	0	0	4507	0.01	0.00	0.00	0	0
1282	0.01	0.00	0.00	0	0	444	0.06	0.00	0.04	0	0	2263	0.00	0.00	0.00	0	0	4511	0.01	0.00	0.00	0	0
2741	0.02	0.00	0.00	0	0	114	0.02	0.00	0.04	0	0	2346	0.00	0.00	0.00	0	0	4714	0.01	0.00	0.00	0	0
99	0.02	0.01	0.06	0	0	124	0.03	0.01	0.06	0	0	1139	0.01	0.00	0.03	0	0	4505	0.02	0.00	0.00	0	0
484	0.04	0.00	0.02	0	0	2094	0.02	0.00	0.01	0	0	423	0.04	0.00	0.02	0	0	3844	0.01	0.00	0.01	0	0
90	0.01	0.00	0.02	0	0	2099	0.03	0.01	0.07	0	0	353	0.06	0.00	0.04	0	0	3843	0.04	0.00	0.00	0	0
398	0.08	0.00	0.04	0	0	6	0.02	0.00	0.03	0	0	1297	0.02	0.00	0.01	0	0	122	0.03	0.01	0.07	0	0
85	0.01	0.00	0.01	0	0	1418	0.05	0.00	0.02	0	0	1301	0.03	0.01	0.07	0	0	91	0.03	0.01	0.08	0	0
1709	0.02	0.01	0.06	0	0	1707	0.02	0.00	0.03	0	0	1076	0.00	0.00	0.00	0	0	997	0.04	0.00	0.01	0	0
5	0.02	0.00	0.02	0	0	1412	0.00	0.00	0.00	0	0	993	0.00	0.00	0.00	0	0	3742	0.01	0.00	0.01	0	0
1456	0.02	0.00	0.03	0	0	1120	0.01	0.00	0.00	0	0	4	0.02	0.00	0.03	0	0	97	0.01	0.00	0.00	0	0
87	0.01	0.00	0.01	0	0	558	0.04	0.00	0.02	0	0	62	0.02	0.00	0.02	0	0	3112	0.04	0.01	0.09	0	0
358	0.01	0.00	0.00	0	0	401	0.01	0.00	0.00	0	0	1074	0.01	0.00	0.00	0	0	735	0.09	0.00	0.03	0	0
1196	0.01	0.00	0.00	0	0	2111	0.01	0.00	0.00	0	0	1317	0.01	0.00	0.01	0	0	103	0.02	0.00	0.03	0	0
1195	0.03	0.00	0.02	0	0	108	0.01	0.00	0.00	0	0	994	0.01	0.00	0.00	0	0	3746	0.01	0.00	0.00	0	0
1704	0.01	0.00	0.01	0	0	113	0.01	0.00	0.01	0	0	319	0.01	0.00	0.00	0	0	95	0.01	0.00	0.01	0	0
1369	0.01	0.00	0.00	0	0							59	0.01	0.00	0.01	0	0	102	0.02	0.00	0.03	0	0
1374	0.01	0.00	0.00	0	0							71	0.04	0.00	0.05	0	0	2679	0.01	0.00	0.00	0	0
1730	0.01	0.00	0.01	0	0							57	0.01	0.00	0.00	0	0	2681	0.03	0.00	0.01	0	0
317	0.00	0.00	0.00	0	0							298	0.00	0.00	0.00	0	0	3759	0.02	0.00	0.03	0	0
																		3	0.04	0.01	0.05	0	0
																		631	0.01	0.00	0.01	0	0
																		3088	0.02	0.00	0.02	0	0

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

▶vec.h: 491 - 1.27 %

ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	ASM Loop ID	Vector Length Use (%)
Run orig_default						Run icx_default						Run aocc_9						Run icx_2
Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.h: 491-497					Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.h: 491-497					Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.h: 491-497					Loop Source Regions
1723	0.10	0.04	0.43	100	75	2109	0.09	0.04	0.42	100	75	1312	0.09	0.04	0.41	100	75

Sum on 1 analyzed binary loop (libggml-cpu.so - 1723)						Sum on 1 analyzed binary loop (libggml-cpu.so - 2109)						Sum on 1 analyzed binary loop (libggml-cpu.so - 1312)						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

▶vec.cpp: 331 - 1.19 %

ASM Loop ID	Vector Length Use (%)	ASM Loop ID	Vector Length Use (%)	ASM Loop ID	Vector Length Use (%)	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)
Run orig_default		Run icx_default		Run aocc_9		Run icx_2
Loop Source Regions		Loop Source Regions		Loop Source Regions		Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/simd-mappings.h: 130-130 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.cpp: 331-332
						1728	0.31	0.16	1.19	5.88	8.82

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.		Sum on 1 analyzed binary loop (libggml-cpu.so - 1728)
Analysis	Count	Analysis	Count	Analysis	Count	Analysis					Count
						Loop Computation Issues
						Less than 10% of the FP ADD/SUB/MUL arithmetic operations are performed using FMA					1
						Data Access Issues
						Presence of indirect access					1
						Vectorization Roadblocks
						Presence of indirect access					1

▶vec.h: 508 - 1.11 %

ASM Loop ID	Vector Length Use (%)	ASM Loop ID	Vector Length Use (%)	ASM Loop ID	Vector Length Use (%)	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)
Run orig_default		Run icx_default		Run aocc_9		Run icx_2
Loop Source Regions		Loop Source Regions		Loop Source Regions		Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/./ggml-impl.h: 346-346 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/./ggml-impl.h: 389-389 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/./ggml-impl.h: 399-404 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.h: 508-509 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/simd-mappings.h: 130-130
						3756	0.27	0.15	1.11	85.05	20.79

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.		Sum on 1 analyzed binary loop (libggml-cpu.so - 3756)
Analysis	Count	Analysis	Count	Analysis	Count	Analysis					Count
						Loop Computation Issues
						Less than 10% of the FP ADD/SUB/MUL arithmetic operations are performed using FMA					1
						Data Access Issues
						Presence of indirect access					1
						Presence of special instructions executing on a single port					1
						Vectorization Roadblocks
						Presence of indirect access					1
						Inefficient Vectorization
						Presence of special instructions executing on a single port					1

▶vec.cpp: 311 - 0.84 %

ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	ASM Loop ID	Vector Length Use (%)
Run orig_default						Run icx_default						Run aocc_9						Run icx_2
Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.cpp: 311-316					Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.cpp: 311-316					Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.cpp: 311-316					Loop Source Regions
761	0.10	0.03	0.28	100	66.67	866	0.08	0.02	0.25	100	66.67	689	0.08	0.03	0.31	100	66.67

Sum on 1 analyzed binary loop (libggml-cpu.so - 761)						Sum on 1 analyzed binary loop (libggml-cpu.so - 866)						Sum on 1 analyzed binary loop (libggml-cpu.so - 689)						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
Loop Computation Issues						Loop Computation Issues						Loop Computation Issues
Low iteration count						Low iteration count					1	Low iteration count
Control Flow Issues						Control Flow Issues						Control Flow Issues
Low iteration count						Low iteration count					1	Low iteration count

▶vec.h: 1084 - 0.52 %

ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	ASM Loop ID	Vector Length Use (%)
Run orig_default						Run icx_default						Run aocc_9						Run icx_2
Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.h: 1084-1101 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.h: 1113-1116 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.cpp: 372-373					Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.h: 1084-1101 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.h: 1113-1116 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.cpp: 372-373					Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.h: 1084-1101 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.h: 1115-1115 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.cpp: 372-373					Loop Source Regions
765	0.22	0.02	0.17	98	98.13	873	0.25	0.02	0.17	100	100	695	0.26	0.02	0.18	98	98.13

Sum on 1 analyzed binary loop (libggml-cpu.so - 765)						Sum on 1 analyzed binary loop (libggml-cpu.so - 873)						Sum on 1 analyzed binary loop (libggml-cpu.so - 695)						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
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
Control Flow Issues						Control Flow Issues						Control Flow Issues
Presence of 2 to 4 paths					1	Presence of 2 to 4 paths					1	Presence of 2 to 4 paths					1
Data Access Issues						Data Access Issues						Data Access Issues
Presence of constant non-unit stride data access					1	Presence of constant non-unit stride data access					0	Presence of constant non-unit stride data access					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					0	Presence of special instructions executing on a single port					1
Vectorization Roadblocks						Vectorization Roadblocks						Vectorization Roadblocks
Presence of 2 to 4 paths					1	Presence of 2 to 4 paths					1	Presence of 2 to 4 paths					1
Presence of constant non-unit stride data access					1	Presence of constant non-unit stride data access					0	Presence of constant non-unit stride data access					1
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					0	Presence of special instructions executing on a single port					1
Use of masked instructions					1	Use of masked instructions					1	Use of masked instructions					1

▶ops.cpp: 6220 - 0.34 %

ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	ASM Loop ID	Vector Length Use (%)	ASM Loop ID	Vector Length Use (%)
Run orig_default						Run icx_default						Run aocc_9		Run icx_2
Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/ops.cpp: 6220-6220 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/ops.cpp: 6229-6230 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/ops.cpp: 6238-6245					Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/ops.cpp: 6220-6220 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/ops.cpp: 6229-6230 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/ops.cpp: 6238-6245					Loop Source Regions		Loop Source Regions
1462	0.05	0.02	0.20	6.67	7.5	1734	0.03	0.01	0.14	15.56	9.17

Sum on 1 analyzed binary loop (libggml-cpu.so - 1462)						Sum on 1 analyzed binary loop (libggml-cpu.so - 1734)						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
Loop Computation Issues						Loop Computation Issues
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					1	Less than 10% of the FP ADD/SUB/MUL arithmetic operations are performed using FMA					1
Presence of a large number of scalar integer instructions					1	Presence of a large number of scalar integer instructions					0
Control Flow Issues						Control Flow Issues
Presence of calls					1	Presence of calls					1
Data Access Issues						Data Access Issues
More than 20% of the loads are accessing the stack					1	More than 20% of the loads are accessing the stack					1
Vectorization Roadblocks						Vectorization Roadblocks
Presence of calls					1	Presence of calls					1

▶ggml-quants.c: 216 - 0.20 %

ASM Loop ID	Vector Length Use (%)	ASM Loop ID	Vector Length Use (%)	ASM Loop ID	Vector Length Use (%)	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)
Run orig_default		Run icx_default		Run aocc_9		Run icx_2
Loop Source Regions		Loop Source Regions		Loop Source Regions		Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-quants.c: 216-219
						726	0.06	0.03	0.20	95	22.5

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.		Sum on 1 analyzed binary loop (libggml-base.so - 726)
Analysis	Count	Analysis	Count	Analysis	Count	Analysis					Count
						Loop Computation Issues
						Less than 10% of the FP ADD/SUB/MUL arithmetic operations are performed using FMA					1
						Presence of a large number of scalar integer instructions					1
						Control Flow Issues
						Presence of calls					1
						Data Access Issues
						More than 10% of the vector loads instructions are unaligned					1
						Presence of special instructions executing on a single port					1
						More than 20% of the loads are accessing the stack					1
						Vectorization Roadblocks
						Presence of calls					1
						Inefficient Vectorization
						Presence of special instructions executing on a single port					1

▶ops.cpp: 6210 - 0.20 %

ASM Loop ID	Vector Length Use (%)	ASM Loop ID	Vector Length Use (%)	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	ASM Loop ID	Vector Length Use (%)
Run orig_default		Run icx_default		Run aocc_9						Run icx_2
Loop Source Regions		Loop Source Regions		Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/ops.cpp: 6210-6211 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/ops.cpp: 6220-6230 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/ops.cpp: 6238-6245					Loop Source Regions
				1142	0.05	0.02	0.20	5.56	7.29

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 - 1142)						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
				Loop Computation Issues
				Presence of expensive FP instructions					1
				Presence of a large number of scalar integer instructions					1
				Control Flow Issues
				Presence of calls					1
				Presence of more than 4 paths					1
				Data Access Issues
				Presence of special instructions executing on a single port					1
				More than 20% of the loads are accessing the stack					1
				Vectorization Roadblocks
				Presence of calls					1
				Presence of more than 4 paths					1
				Inefficient Vectorization
				Presence of special instructions executing on a single port					1
				Use of masked instructions					1

▶vec.h: 1190 - 0.15 %

ASM Loop ID	Vector Length Use (%)	ASM Loop ID	Vector Length Use (%)	ASM Loop ID	Vector Length Use (%)	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)
Run orig_default		Run icx_default		Run aocc_9		Run icx_2
Loop Source Regions		Loop Source Regions		Loop Source Regions		Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.h: 1190-1214 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.h: 1222-1225 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.cpp: 380-381
						1742	0.26	0.02	0.15	98.03	24.75

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.		Sum on 1 analyzed binary loop (libggml-cpu.so - 1742)
Analysis	Count	Analysis	Count	Analysis	Count	Analysis					Count
						Loop Computation Issues
						Presence of expensive FP instructions					1
						Less than 10% of the FP ADD/SUB/MUL arithmetic operations are performed using FMA					1
						Control Flow Issues
						Presence of 2 to 4 paths					1
						Data Access Issues
						More than 10% of the vector loads instructions are unaligned					1
						Vectorization Roadblocks
						Presence of 2 to 4 paths					1

▶sgemm.cpp: 205 - 0.11 %

ASM Loop ID	Vector Length Use (%)	ASM Loop ID	Vector Length Use (%)	ASM Loop ID	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	ASM Loop ID	Vector Length Use (%)
Run orig_default		Run icx_default		Run aocc_9						Run icx_2
Loop Source Regions		Loop Source Regions		Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/llamafile/sgemm.cpp: 205-206 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/llamafile/sgemm.cpp: 220-221 /beegfs/hackathon/users/eoseret/qaas_runs_test/gmz17.benchmarkcenter.megware.com/176-976-8905/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/llamafile/sgemm.cpp: 851-853					Loop Source Regions
				1812	0.03	0.01	0.11	71.43	17.86

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 - 1812)						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
				Loop Computation Issues
				Less than 10% of the FP ADD/SUB/MUL arithmetic operations are performed using FMA					1
				Data Access Issues
				Presence of constant non-unit stride data access					1
				Presence of special instructions executing on a single port					1
				Vectorization Roadblocks
				Presence of constant non-unit stride data access					1
				Inefficient Vectorization
				Presence of special instructions executing on a single port					1

Report Configuration

Loops

▶sgemm.cpp: 138 - 163.87 %

▶quants.c: 1066 - 78.11 %

▶quants.c: 298 - 1.71 %

▶<unknown>: 0 - 1.50 %

▶vec.h: 491 - 1.27 %

▶vec.cpp: 331 - 1.19 %

▶vec.h: 508 - 1.11 %

▶vec.cpp: 311 - 0.84 %

▶vec.h: 1084 - 0.52 %

▶ops.cpp: 6220 - 0.34 %

▶ggml-quants.c: 216 - 0.20 %

▶ops.cpp: 6210 - 0.20 %

▶vec.h: 1190 - 0.15 %

▶sgemm.cpp: 205 - 0.11 %