OV - Compare Loops

Loops

▶mmq.cpp: 1138 - 14.95 %

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 aocc_default		Run icx_4		Run aocc_2
Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1138-1151 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 2035-2040 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 2060-2074 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 2093-2101 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 2146-2152 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 2488-2488 /usr/lib/gcc/x86_64-redhat-linux/11/../../../../include/c++/11/bits/stl_algobase.h: 235-235						Loop Source Regions		Loop Source Regions		Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1138-1151 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 2035-2040 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 2060-2079 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 2085-2096 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 2146-2152 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 2488-2488 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 2057-2057 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 2101-2101 /usr/lib/gcc/x86_64-redhat-linux/11/../../../../include/c++/11/bits/stl_algobase.h: 235-235
565	0.01	0.00	0.18	0	0	242.33					695	0.01	0.00	0.22	0	0	318.49
577	0.20	0.14	14.17	0	0	488.56					715	0.02	0.00	0.39	0	0	128.41

Sum on 2 analyzed binary loops (libggml-cpu.so - 565, libggml-cpu.so - 577)							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 2 analyzed binary loops (libggml-cpu.so - 695, libggml-cpu.so - 715)
Analysis						Count	Analysis	Count	Analysis	Count	Analysis						Count
Control Flow Issues											Control Flow Issues
Vectorization Roadblocks											Vectorization Roadblocks
Presence of more than 4 paths						1					Presence of more than 4 paths						1

▶mmq.cpp: 303 - 12.63 %

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 aocc_default							Run icx_4							Run aocc_2
Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 303-330 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 651-663 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 816-818 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 825-825 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1390-1392						Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 303-330 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 651-663 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 816-818 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 825-825 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1390-1392						Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 303-330 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 651-663 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 816-818 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 825-825 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1390-1392						Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 303-330 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 651-663 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 816-818 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 825-825 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1390-1392
659	0.05	0.03	2.97	90.91	38.76	0	384	0.06	0.03	3.23	90.91	38.76	0	675	0.06	0.03	3.24	90.91	38.76	0	384	0.05	0.03	3.19	90.91	38.76	0

Sum on 1 analyzed binary loop (libggml-cpu.so - 659)							Sum on 1 analyzed binary loop (libggml-cpu.so - 384)							Sum on 1 analyzed binary loop (libggml-cpu.so - 675)							Sum on 1 analyzed binary loop (libggml-cpu.so - 384)
Analysis						Count	Analysis						Count	Analysis						Count	Analysis						Count
Loop Computation Issues							Loop Computation Issues							Loop Computation Issues							Loop Computation Issues
Presence of a large number of scalar integer instructions						1	Presence of a large number of scalar integer instructions						1	Presence of a large number of scalar integer instructions						1	Presence of a large number of scalar integer instructions						1
Data Access Issues							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						1	Presence of constant non-unit stride data access						1	Presence of constant non-unit stride data access						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
More than 20% of the loads are accessing the stack						1	More than 20% of the loads are accessing the stack						1	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							Vectorization Roadblocks							Vectorization Roadblocks
Presence of constant non-unit stride data access						1	Presence of constant non-unit stride data access						1	Presence of constant non-unit stride data access						1	Presence of constant non-unit stride data access						1
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

▶<unknown>: 0 - 2.34 %

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 aocc_default							Run icx_4							Run aocc_2
Loop Source Regions							Loop Source Regions							Loop Source Regions							Loop Source Regions
2392	0.00	0.00	0.01	0	0	0	2749	0.00	0.00	0.01	0	0	0	2412	0.00	0.00	0.01	0	0	0	3042	0.00	0.00	0.01	0	0	0
2504	0.01	0.00	0.01	0	0	0	2741	0.00	0.00	0.01	0	0	0	2417	0.01	0.00	0.02	0	0	0	2934	0.01	0.00	0.01	0	0	0
2395	0.00	0.00	0.01	0	0	0	2640	0.01	0.00	0.01	0	0	0	2414	0.01	0.00	0.01	0	0	0	2820	0.00	0.00	0.01	0	0	0
2294	0.02	0.00	0.02	0	0	0	2628	0.01	0.00	0.02	0	0	0	2524	0.00	0.00	0.01	0	0	0	2931	0.00	0.00	0.01	0	0	0
2289	0.01	0.00	0.01	0	0	0	2842	0.01	0.00	0.01	0	0	0	2521	0.01	0.00	0.01	0	0	0	2832	0.01	0.00	0.01	0	0	0
2298	0.02	0.00	0.02	0	0	0	2638	0.02	0.00	0.02	0	0	0	2275	0.00	0.00	0.01	0	0	0	2638	0.01	0.00	0.01	0	0	0
1955	0.00	0.00	0.03	0	0	0	2738	0.01	0.00	0.01	0	0	0	2638	0.01	0.00	0.01	0	0	0	703	0.00	0.00	0.00	0	0	NA
1610	0.01	0.00	0.08	0	0	88.47	816	0.01	0.00	0.05	0	0	67.91	2421	0.02	0.00	0.02	0	0	0	3377	0.08	0.00	0.09	0	0	435.17
1140	0.01	0.00	0.05	0	0	576.2	695	0.00	0.00	0.01	0	0	583.81	2594	0.00	0.00	0.02	0	0	2.25	812	0.00	0.00	0.07	0	0	34.94
1767	0.00	0.00	0.02	0	0	0	107	0.00	0.00	0.00	0	0	NA	752	0.00	0.00	0.05	0	0	65.69	1288	0.01	0.00	0.09	0	0	379.12
3016	0.08	0.00	0.08	0	0	455.94	2441	0.00	0.00	0.00	0	0	NA	2048	0.00	0.00	0.01	0	0	0	407	0.00	0.00	0.00	0	0	NA
6	0.01	0.00	0.09	0	0	0	2105	0.00	0.00	0.01	0	0	711.52	1305	0.01	0.00	0.05	0	0	464.32	107	0.00	0.00	0.01	0	0	27.34
2250	0.00	0.00	0.00	0	0	NA	914	0.00	0.00	0.05	0	0	100.34	149	0.00	0.00	0.02	0	0	0	2459	0.00	0.00	0.01	0	0	27.34
575	0.00	0.00	0.00	0	0	NA	5	0.00	0.00	0.08	0	0	0	6	0.01	0.00	0.09	0	0	0	2454	0.00	0.00	0.00	0	0	NA
2265	0.00	0.00	0.01	0	0	17.95	1759	0.00	0.00	0.01	0	0	437.86	924	0.01	0.00	0.10	0	0	40.49	2160	0.01	0.00	0.05	0	0	152.65
146	0.00	0.00	0.02	0	0	0	701	0.00	0.00	0.01	0	0	291.91	2608	0.00	0.00	0.00	0	0	NA	1838	0.00	0.00	0.01	0	0	1321.46
1605	0.00	0.00	0.01	0	0	215.4	407	0.00	0.00	0.01	0	0	0	433	0.00	0.00	0.00	0	0	NA	108	0.00	0.00	0.02	0	0	0
2254	0.00	0.00	0.02	0	0	2.99	1969	0.00	0.00	0.01	0	0	0	2152	0.00	0.00	0.01	0	0	1124.79	704	0.00	0.00	0.01	0	0	437.45
833	0.00	0.00	0.07	0	0	62.08	2419	0.00	0.00	0.04	0	0	1.3	1885	0.01	0.00	0.06	0	0	105.73	2168	0.00	0.00	0.09	0	0	123.94
719	0.00	0.00	0.06	0	0	59.56	1758	0.01	0.00	0.08	0	0	87.96	3406	0.07	0.00	0.07	0	0	489.77	1837	0.00	0.00	0.06	0	0	0
1863	0.00	0.00	0.01	0	0	543	1232	0.00	0.00	0.05	0	0	570.64	584	0.00	0.00	0.00	0	0	NA	828	0.00	0.00	0.01	0	0	0
2263	0.01	0.00	0.09	0	0	301.79	317	0.00	0.00	0.01	0	0	0	1880	0.00	0.00	0.00	0	0	NA	816	0.00	0.00	0.01	0	0	0
														572	0.00	0.00	0.01	0	0	0
														865	0.00	0.00	0.01	0	0	0
														1971	0.00	0.00	0.02	0	0	0
														194	0.00	0.00	0.01	0	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

▶mmq.cpp: 1140 - 2.16 %

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 aocc_default							Run icx_4							Run aocc_2
Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1140-1142 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1150-1151						Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1140-1142 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1150-1151						Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1140-1142 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1150-1151						Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1140-1142 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1150-1151
567	0.02	0.00	0.28	100	100	1940.04	705	0.01	0.00	0.16	100	100	3018.11	578	0.01	0.00	0.29	100	100	1831.67	696	0.01	0.00	0.20	100	100	591.84
566	0.02	0.00	0.47	100	100	287.2	704	0.01	0.00	0.21	100	100	518.95	577	0.03	0.00	0.38	100	100	306.2	697	0.01	0.00	0.17	100	100	3042.02

Sum on 2 analyzed binary loops (libggml-cpu.so - 567, libggml-cpu.so - 566)							Sum on 2 analyzed binary loops (libggml-cpu.so - 705, libggml-cpu.so - 704)							Sum on 2 analyzed binary loops (libggml-cpu.so - 578, libggml-cpu.so - 577)							Sum on 2 analyzed binary loops (libggml-cpu.so - 696, libggml-cpu.so - 697)
Analysis						Count	Analysis						Count	Analysis						Count	Analysis						Count
Loop Computation Issues							Loop Computation Issues							Loop Computation Issues							Loop Computation Issues
Presence of a large number of scalar integer instructions						1	Presence of a large number of scalar integer instructions							Presence of a large number of scalar integer instructions						1	Presence of a large number of scalar integer instructions
Low iteration count						0	Low iteration count							Low iteration count						1	Low iteration count
Control Flow Issues							Control Flow Issues							Control Flow Issues							Control Flow Issues
Low iteration count							Low iteration count							Low iteration count						1	Low iteration count
Data Access Issues							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						1	Presence of constant non-unit stride data access							Presence of constant non-unit stride data access						1
Presence of indirect access						1	Presence of indirect access						1	Presence of indirect access							Presence of indirect access						1
Vectorization Roadblocks							Vectorization Roadblocks							Vectorization Roadblocks							Vectorization Roadblocks
Presence of more than 4 paths						0	Presence of more than 4 paths						0	Presence of more than 4 paths						1	Presence of more than 4 paths						0
Presence of constant non-unit stride data access						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
Presence of indirect access						1	Presence of indirect access						1	Presence of indirect access						0	Presence of indirect access						1

▶vec.h: 1084 - 1.09 %

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 aocc_default							Run icx_4							Run aocc_2
Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.cpp: 372-373 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.h: 1084-1101 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.h: 1113-1116						Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.cpp: 372-373 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.h: 1084-1101 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.h: 1113-1116						Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.cpp: 372-373 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.h: 1084-1101 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.h: 1113-1116						Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.cpp: 372-373 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.h: 1084-1101 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.h: 1115-1115
1147	0.03	0.00	0.34	100	100	324.26	1236	0.02	0.00	0.32	98	98.13	368.86	1312	0.01	0.00	0.20	100	100	544.77	1298	0.01	0.00	0.23	98	98.13	696.27

Sum on 1 analyzed binary loop (libggml-cpu.so - 1147)							Sum on 1 analyzed binary loop (libggml-cpu.so - 1236)							Sum on 1 analyzed binary loop (libggml-cpu.so - 1312)							Sum on 1 analyzed binary loop (libggml-cpu.so - 1298)
Analysis						Count	Analysis						Count	Analysis						Count	Analysis						Count
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
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						1	Presence of 2 to 4 paths						1	Presence of 2 to 4 paths						1
Data Access Issues							Data Access Issues							Data Access Issues							Data Access Issues
Presence of constant non-unit stride data access							Presence of constant non-unit stride data access						1	Presence of constant non-unit stride data access							Presence of constant non-unit stride data access						1
Presence of special instructions executing on a single port							Presence of special instructions executing on a single port						1	Presence of special instructions executing on a single port							Presence of special instructions executing on a single port						1
Vectorization Roadblocks							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 2 to 4 paths						1
Presence of constant non-unit stride data access						0	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							Inefficient Vectorization
Presence of special instructions executing on a single port						0	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	Use of masked instructions						1

▶ops.cpp: 6220 - 0.40 %

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 aocc_default							Run icx_4							Run aocc_2
Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/ops.cpp: 6220-6222 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/ops.cpp: 6229-6230 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/ops.cpp: 6238-6245						Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/ops.cpp: 6220-6220 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/ops.cpp: 6229-6230 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/ops.cpp: 6238-6245						Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/ops.cpp: 6220-6222 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/ops.cpp: 6229-6230 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/ops.cpp: 6238-6245						Loop Source Regions
1888	0.01	0.00	0.16	15.56	9.17	205.83	2113	0.01	0.00	0.12	1.96	6.62	422.81	2185	0.01	0.00	0.12	15.56	9.17	304.77

Sum on 1 analyzed binary loop (libggml-cpu.so - 1888)							Sum on 1 analyzed binary loop (libggml-cpu.so - 2113)							Sum on 1 analyzed binary loop (libggml-cpu.so - 2185)							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						1	Less than 10% of the FP ADD/SUB/MUL arithmetic operations are performed using FMA						1
Control Flow Issues							Control Flow Issues							Control Flow Issues
Presence of calls						1	Presence of calls						1	Presence of calls						1
Data Access Issues							Data Access Issues							Data Access Issues
Presence of constant non-unit stride data access						0	Presence of constant non-unit stride data access						0	Presence of constant non-unit stride data access						1
More than 20% of the loads are accessing the stack						1	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							Vectorization Roadblocks
Presence of calls						1	Presence of calls						1	Presence of calls						1
Presence of constant non-unit stride data access						0	Presence of constant non-unit stride data access						0	Presence of constant non-unit stride data access						1

▶mmq.cpp: 1570 - 0.40 %

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	Max Time Over Threads (s)	Time w.r.t. Wall Time (s)	Cov (%)	Vect. Ratio (%)	Vector Length Use (%)	GFLOP/s
Run orig_default		Run aocc_default							Run icx_4		Run aocc_2
Loop Source Regions		Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1570-1575 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1584-1584 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1590-1597						Loop Source Regions		Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1570-1575 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1584-1584 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1590-1597
		531	0.00	0.00	0.18	0	0	14.12			531	0.00	0.00	0.21	0	0	35.47

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 - 531)							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 - 531)
Analysis	Count	Analysis						Count	Analysis	Count	Analysis						Count

▶vec.h: 491 - 0.39 %

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	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 aocc_default							Run icx_4							Run aocc_2
Loop Source Regions		Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.h: 491-497						Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.h: 491-497						Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/vec.h: 491-497
		2437	0.01	0.00	0.15	0	0	273.31	2605	0.01	0.00	0.13	0	0	204.8	2477	0.01	0.00	0.10	0	0	334.16

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 - 2437)							Sum on 1 analyzed binary loop (libggml-cpu.so - 2605)							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
		Control Flow Issues							Control Flow Issues
		Presence of 2 to 4 paths						1	Presence of 2 to 4 paths
		Vectorization Roadblocks							Vectorization Roadblocks
		Presence of 2 to 4 paths						1	Presence of 2 to 4 paths

▶mmq.cpp: 1573 - 0.32 %

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	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 aocc_default		Run icx_4							Run aocc_2
Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1573-1575 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1584-1584 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1590-1597						Loop Source Regions		Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1573-1575 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1584-1584 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 1590-1597						Loop Source Regions
475	0.00	0.00	0.17	0	0	13.9			474	0.00	0.00	0.15	0	0	14.89

Sum on 1 analyzed binary loop (libggml-cpu.so - 475)							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 - 474)							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

▶mmq.cpp: 2068 - 0.15 %

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 aocc_default							Run icx_4		Run aocc_2
Loop Source Regions		Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/amx/mmq.cpp: 2068-2078						Loop Source Regions		Loop Source Regions
		710	0.01	0.00	0.15	0	0	0

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 - 710)							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

▶binary-ops.cpp: 18 - 0.13 %

ASM Loop ID	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 aocc_default		Run icx_4		Run aocc_2
Loop Source Regions		Loop Source Regions		Loop Source Regions		Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/binary-ops.cpp: 18-18 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/binary-ops.cpp: 31-32
						910	0.01	0.00	0.13	0	6.25	24.57

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 - 910)
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

▶quants.c: 298 - 0.12 %

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 aocc_default							Run icx_4		Run aocc_2
Loop Source Regions		Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/arch/x86/quants.c: 298-347 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/arch/x86/quants.c: 353-355 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/./ggml-impl.h: 346-346 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/./ggml-impl.h: 355-355 /beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/./ggml-impl.h: 389-404						Loop Source Regions		Loop Source Regions
		3307	0.10	0.00	0.12	58.33	28.75	334.48

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 - 3307)							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
		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 more than 4 paths						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
		Vectorization Roadblocks
		Presence of more than 4 paths						1
		Inefficient Vectorization
		Presence of special instructions executing on a single port						1

▶ggml-cpu.c: 3204 - 0.10 %

ASM Loop ID	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 aocc_default		Run icx_4		Run aocc_2
Loop Source Regions		Loop Source Regions		Loop Source Regions		Loop Source Regions	/beegfs/hackathon/users/eoseret/qaas_runs_test/isix06.benchmarkcenter.megware.com/177-002-9891/llama.cpp/build/llama.cpp/ggml/src/ggml-cpu/ggml-cpu.c: 3204-3207
						5	0.02	0.00	0.10	100	100	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

Report Configuration

Loops

▶mmq.cpp: 1138 - 14.95 %

▶mmq.cpp: 303 - 12.63 %

▶<unknown>: 0 - 2.34 %

▶mmq.cpp: 1140 - 2.16 %

▶vec.h: 1084 - 1.09 %

▶ops.cpp: 6220 - 0.40 %

▶mmq.cpp: 1570 - 0.40 %

▶vec.h: 491 - 0.39 %

▶mmq.cpp: 1573 - 0.32 %

▶mmq.cpp: 2068 - 0.15 %

▶binary-ops.cpp: 18 - 0.13 %

▶quants.c: 298 - 0.12 %

▶ggml-cpu.c: 3204 - 0.10 %