llama.cpp/tests/fusion/MTL.csv at b11294

mirror of https://github.com/ggml-org/llama.cpp.git synced 2026-10-01 02:17:42 -05:00

Files

T

Pascal db00347a4b ci : fix Fusion / metal by adding glm5-next to MTL.csv (#29712 )

#27773 adds the glm5-next arch without its rows in the Metal fusion
baseline, so test-fusion --check fails on it. The rows come from
test-fusion --record on an M5 Max, and --check passes 270/270.

2026-09-30 09:50:23 +02:00

19 KiB

Raw Permalink Blame History

1	# test-fusion baseline for device MTL
2	# arch ,moe ,mode ,label , count
3	afmoe ,1 ,any ,MUL+ADD , 2
4	afmoe ,1 ,any ,RMS_NORM+MUL , 10
5	afmoe ,1 ,any ,RMS_NORM+MUL+ADD , 3
6	arcee ,0 ,any ,RMS_NORM+MUL , 5
7	arctic ,0 ,any ,MUL+ADD , 4
8	arctic ,0 ,any ,RMS_NORM+MUL , 7
9	arctic ,0 ,any ,SOFT_MAX+ARGSORT+GET_ROWS+SUM_ROWS+CLAMP+DIV, 2
10	baichuan ,0 ,any ,RMS_NORM+MUL , 5
11	bailingmoe ,1 ,any ,ADD+ADD , 2
12	bailingmoe ,1 ,any ,MUL+ADD , 4
13	bailingmoe ,1 ,any ,RMS_NORM+MUL , 5
14	bailingmoe ,1 ,any ,SOFT_MAX+ARGSORT+GET_ROWS , 2
15	bailingmoe2 ,1 ,any ,ADD+ADD , 1
16	bailingmoe2 ,1 ,any ,MUL+ADD , 2
17	bailingmoe2 ,1 ,any ,RMS_NORM+MUL , 9
18	bailingmoe3 ,1 ,any ,ADD+ADD , 1
19	bailingmoe3 ,1 ,any ,GATED_DELTA_NET+CPY , 1
20	bailingmoe3 ,1 ,any ,MUL+ADD , 2
21	bailingmoe3 ,1 ,any ,RMS_NORM+MUL , 8
22	bailingmoe3 ,1 ,any ,RMS_NORM+SCALE , 2
23	bloom ,0 ,any ,NORM+MUL+ADD , 6
24	chatglm ,0 ,any ,RMS_NORM+MUL , 5
25	codeshell ,0 ,any ,NORM+MUL+ADD , 5
26	cogvlm ,0 ,any ,RMS_NORM+MUL , 5
27	cohere2 ,0 ,any ,ADD+ADD , 2
28	cohere2 ,0 ,any ,NORM+MUL , 3
29	cohere2moe ,1 ,any ,ADD+ADD , 2
30	cohere2moe ,1 ,any ,MUL+ADD , 2
31	cohere2moe ,1 ,any ,RMS_NORM+MUL , 3
32	command-r ,0 ,any ,NORM+MUL , 3
33	dbrx ,0 ,any ,MUL+ADD , 4
34	dbrx ,0 ,any ,NORM+MUL , 5
35	dbrx ,0 ,any ,SOFT_MAX+ARGSORT+GET_ROWS+SUM_ROWS+CLAMP+DIV, 2
36	deci ,0 ,any ,RMS_NORM+MUL , 5
37	deepseek ,0 ,any ,ADD+ADD , 1
38	deepseek ,0 ,any ,MUL+ADD , 2
39	deepseek ,0 ,any ,RMS_NORM+MUL , 5
40	deepseek ,0 ,any ,SOFT_MAX+ARGSORT+GET_ROWS , 1
41	deepseek2 ,0 ,any ,ADD+ADD , 1
42	deepseek2 ,0 ,any ,MUL+ADD , 2
43	deepseek2 ,0 ,any ,RMS_NORM+MUL , 9
44	deepseek32 ,0 ,any ,ADD+ADD , 1
45	deepseek32 ,0 ,any ,MUL+ADD , 2
46	deepseek32 ,0 ,any ,NORM+MUL+ADD , 2
47	deepseek32 ,0 ,any ,RMS_NORM+MUL , 9
48	deepseek4 ,0 ,any ,MUL+ADD , 8
49	deepseek4 ,0 ,any ,RMS_NORM+MUL , 20
50	dots1 ,0 ,any ,ADD+ADD , 1
51	dots1 ,0 ,any ,MUL+ADD , 2
52	dots1 ,0 ,any ,RMS_NORM+MUL , 9
53	dots3note ,0 ,any ,ADD+ADD , 1
54	dots3note ,0 ,any ,MUL+ADD , 2
55	dots3note ,0 ,any ,NORM+MUL+ADD , 1
56	dots3note ,0 ,any ,RMS_NORM+MUL , 11
57	dream ,0 ,any ,RMS_NORM+MUL , 5
58	ernie4_5-moe ,1 ,any ,ADD+ADD , 1
59	ernie4_5-moe ,1 ,any ,MUL+ADD , 2
60	ernie4_5-moe ,1 ,any ,RMS_NORM+MUL , 5
61	ernie4_5 ,0 ,any ,RMS_NORM+MUL , 5
62	exaone ,0 ,any ,RMS_NORM+MUL , 5
63	exaone-moe ,1 ,any ,ADD+ADD , 1
64	exaone-moe ,1 ,any ,MUL+ADD , 2
65	exaone-moe ,1 ,any ,RMS_NORM+MUL , 9
66	exaone4 ,0 ,any ,RMS_NORM+MUL , 5
67	exaone4 ,0 ,any ,RMS_NORM+MUL+ADD , 4
68	falcon ,0 ,any ,ADD+ADD , 2
69	falcon ,0 ,any ,NORM+MUL+ADD , 5
70	falcon-h1 ,0 ,any ,ADD+ADD , 2
71	falcon-h1 ,0 ,any ,RMS_NORM+MUL , 9
72	gemma ,0 ,any ,RMS_NORM+MUL , 5
73	gemma2 ,0 ,any ,RMS_NORM+MUL , 5
74	gemma2 ,0 ,any ,RMS_NORM+MUL+ADD , 4
75	gemma3 ,0 ,any ,RMS_NORM+MUL , 9
76	gemma3 ,0 ,any ,RMS_NORM+MUL+ADD , 4
77	glm-dsa ,0 ,any ,ADD+ADD , 1
78	glm-dsa ,0 ,any ,MUL+ADD , 2
79	glm-dsa ,0 ,any ,NORM+MUL+ADD , 2
80	glm-dsa ,0 ,any ,RMS_NORM+MUL , 9
81	glm4 ,0 ,any ,RMS_NORM+MUL , 5
82	glm4 ,0 ,any ,RMS_NORM+MUL+ADD , 4
83	glm4moe ,1 ,any ,ADD+ADD , 1
84	glm4moe ,1 ,any ,MUL+ADD , 2
85	glm4moe ,1 ,any ,RMS_NORM+MUL , 9
86	glm5-next ,0 ,any ,ADD+ADD+ADD , 1
87	glm5-next ,0 ,any ,GATED_DELTA_NET+CPY , 1
88	glm5-next ,0 ,any ,MUL+ADD , 2
89	glm5-next ,0 ,any ,NORM+MUL+ADD , 1
90	glm5-next ,0 ,any ,RMS_NORM+MUL , 8
91	glm5-next ,0 ,any ,RMS_NORM+SCALE , 2
92	gpt-oss ,0 ,any ,MUL+ADD , 4
93	gpt-oss ,0 ,any ,RMS_NORM+MUL , 5
94	gpt2 ,0 ,any ,NORM+MUL+ADD , 5
95	gptneox ,0 ,any ,NORM+MUL+ADD , 5
96	granite ,0 ,any ,RMS_NORM+MUL , 5
97	granite ,0 ,any ,MUL+ADD , 4
98	granite ,0 ,any ,RMS_NORM+MUL , 5
99	granite ,0 ,any ,SOFT_MAX+ARGSORT+GET_ROWS+SUM_ROWS+CLAMP+DIV, 2
100	granite_swa ,0 ,any ,RMS_NORM+MUL , 5
101	granitehybrid ,0 ,any ,RMS_NORM+MUL , 6
102	granitemoe ,1 ,any ,RMS_NORM+MUL , 5
103	granitemoe ,1 ,any ,MUL+ADD , 4
104	granitemoe ,1 ,any ,RMS_NORM+MUL , 5
105	granitemoe ,1 ,any ,SOFT_MAX+ARGSORT+GET_ROWS+SUM_ROWS+CLAMP+DIV, 2
106	grok ,0 ,any ,MUL+ADD , 4
107	grok ,0 ,any ,RMS_NORM+MUL , 5
108	grok ,0 ,any ,RMS_NORM+MUL+ADD , 4
109	grok ,0 ,any ,SOFT_MAX+ARGSORT+GET_ROWS+SUM_ROWS+CLAMP+DIV, 2
110	grovemoe ,1 ,any ,ADD+ADD , 2
111	grovemoe ,1 ,any ,MUL+ADD , 8
112	grovemoe ,1 ,any ,RMS_NORM+MUL , 9
113	hunyuan-dense ,0 ,any ,RMS_NORM+MUL , 9
114	hunyuan-moe ,1 ,any ,ADD+ADD , 2
115	hunyuan-moe ,1 ,any ,MUL+ADD , 4
116	hunyuan-moe ,1 ,any ,RMS_NORM+MUL , 9
117	hunyuan-moe ,1 ,any ,SOFT_MAX+ARGSORT+GET_ROWS+SUM_ROWS+CLAMP+DIV, 2
118	hunyuan_vl ,0 ,any ,RMS_NORM+MUL , 9
119	hy_v3 ,0 ,any ,ADD+ADD , 2
120	hy_v3 ,0 ,any ,MUL+ADD , 4
121	hy_v3 ,0 ,any ,RMS_NORM+MUL , 9
122	hy_v4 ,0 ,any ,MUL+ADD , 2
123	hy_v4 ,0 ,any ,NORM+MUL+ADD , 1
124	hy_v4 ,0 ,any ,RMS_NORM+MUL , 9
125	internlm2 ,0 ,any ,RMS_NORM+MUL , 5
126	jais ,0 ,any ,NORM+MUL+ADD , 5
127	jais2 ,0 ,any ,NORM+MUL+ADD , 5
128	jamba ,0 ,any ,RMS_NORM+MUL , 8
129	kimi-k3 ,0 ,any ,GATED_DELTA_NET+CPY , 1
130	kimi-k3 ,0 ,any ,MUL+ADD , 2
131	kimi-k3 ,0 ,any ,RMS_NORM+MUL , 17
132	kimi-k3 ,0 ,any ,RMS_NORM+SCALE , 2
133	kimi-linear ,0 ,any ,ADD+ADD , 1
134	kimi-linear ,0 ,any ,GATED_DELTA_NET+CPY , 1
135	kimi-linear ,0 ,any ,MUL+ADD , 2
136	kimi-linear ,0 ,any ,RMS_NORM+MUL , 7
137	kimi-linear ,0 ,any ,RMS_NORM+SCALE , 2
138	laguna ,0 ,any ,ADD+ADD , 1
139	laguna ,0 ,any ,MUL+ADD , 2
140	laguna ,0 ,any ,RMS_NORM+MUL , 9
141	lfm2 ,0 ,any ,RMS_NORM+MUL , 7
142	lfm2moe ,1 ,any ,MUL+ADD , 2
143	lfm2moe ,1 ,any ,RMS_NORM+MUL , 7
144	llada ,0 ,any ,RMS_NORM+MUL , 5
145	llada-moe ,1 ,any ,MUL+ADD , 4
146	llada-moe ,1 ,any ,RMS_NORM+MUL , 9
147	llada-moe ,1 ,any ,SOFT_MAX+ARGSORT+GET_ROWS , 2
148	llama ,0 ,any ,RMS_NORM+MUL , 5
149	llama ,0 ,any ,MUL+ADD , 4
150	llama ,0 ,any ,RMS_NORM+MUL , 5
151	llama ,0 ,any ,SOFT_MAX+ARGSORT+GET_ROWS+SUM_ROWS+CLAMP+DIV, 2
152	llama4 ,0 ,any ,ADD+ADD , 2
153	llama4 ,0 ,any ,RMS_NORM+MUL , 9
154	maincoder ,0 ,any ,RMS_NORM+MUL , 9
155	mamba ,0 ,any ,RMS_NORM+MUL , 3
156	mamba2 ,0 ,any ,RMS_NORM+MUL , 5
157	maple ,0 ,any ,MUL+ADD , 4
158	maple ,0 ,any ,RMS_NORM+MUL , 9
159	maple ,0 ,any ,SOFT_MAX+ARGSORT+GET_ROWS+SUM_ROWS+CLAMP+DIV, 2
160	mellum ,0 ,any ,MUL+ADD , 4
161	mellum ,0 ,any ,RMS_NORM+MUL , 9
162	mellum ,0 ,any ,SOFT_MAX+ARGSORT+GET_ROWS+SUM_ROWS+CLAMP+DIV, 2
163	mimo2 ,0 ,any ,MUL+ADD , 4
164	mimo2 ,0 ,any ,RMS_NORM+MUL , 5
165	minicpm ,0 ,any ,RMS_NORM+MUL , 5
166	minicpm ,0 ,any ,MUL+ADD , 4
167	minicpm ,0 ,any ,RMS_NORM+MUL , 5
168	minicpm ,0 ,any ,SOFT_MAX+ARGSORT+GET_ROWS+SUM_ROWS+CLAMP+DIV, 2
169	minicpm3 ,0 ,any ,RMS_NORM+MUL , 9
170	minimax-01 ,0 ,any ,MUL+ADD , 4
171	minimax-01 ,0 ,any ,RMS_NORM+MUL , 6
172	minimax-01 ,0 ,any ,SOFT_MAX+ARGSORT+GET_ROWS+SUM_ROWS+CLAMP+DIV, 2
173	minimax-m2 ,0 ,any ,MUL+ADD , 4
174	minimax-m2 ,0 ,any ,RMS_NORM+MUL , 9
175	minimax-m3 ,0 ,any ,ADD+ADD , 1
176	minimax-m3 ,0 ,any ,MUL+ADD , 2
177	minimax-m3 ,0 ,any ,RMS_NORM+MUL , 11
178	mistral3 ,0 ,any ,RMS_NORM+MUL , 5
179	mistral3 ,0 ,any ,MUL+ADD , 4
180	mistral3 ,0 ,any ,RMS_NORM+MUL , 5
181	mistral3 ,0 ,any ,SOFT_MAX+ARGSORT+GET_ROWS+SUM_ROWS+CLAMP+DIV, 2
182	mistral4 ,0 ,any ,ADD+ADD , 1
183	mistral4 ,0 ,any ,MUL+ADD , 2
184	mistral4 ,0 ,any ,RMS_NORM+MUL , 9
185	mpt ,0 ,any ,NORM+MUL+ADD , 5
186	muse-glimmer ,0 ,any ,RMS_NORM+MUL , 10
187	muse-glimmer ,0 ,any ,RMS_NORM+MUL+ADD , 3
188	nanbeige ,0 ,any ,RMS_NORM+MUL , 5
189	nemotron ,0 ,any ,NORM+MUL+ADD , 5
190	nemotron_h ,0 ,any ,RMS_NORM+MUL , 5
191	nemotron_h_moe ,1 ,any ,RMS_NORM+MUL , 5
192	olmo2 ,0 ,any ,RMS_NORM+MUL , 5
193	olmo2 ,0 ,any ,RMS_NORM+MUL+ADD , 4
194	olmoe ,1 ,any ,MUL+ADD , 4
195	olmoe ,1 ,any ,RMS_NORM+MUL , 9
196	olmoe ,1 ,any ,SOFT_MAX+ARGSORT+GET_ROWS , 2
197	openelm ,0 ,any ,RMS_NORM+MUL , 9
198	orion ,0 ,any ,NORM+MUL+ADD , 5
199	paddleocr ,0 ,any ,RMS_NORM+MUL , 5
200	pangu-embedded ,0 ,any ,RMS_NORM+MUL , 5
201	phi2 ,0 ,any ,ADD+ADD , 2
202	phi2 ,0 ,any ,NORM+MUL+ADD , 3
203	phi3 ,0 ,any ,RMS_NORM+MUL , 5
204	phimoe ,1 ,any ,MUL+ADD , 4
205	phimoe ,1 ,any ,RMS_NORM+MUL+ADD , 5
206	phimoe ,1 ,any ,SOFT_MAX+ARGSORT+GET_ROWS+SUM_ROWS+CLAMP+DIV, 2
207	plamo ,0 ,any ,ADD+ADD , 2
208	plamo ,0 ,any ,RMS_NORM+MUL , 3
209	plamo2 ,0 ,any ,RMS_NORM+MUL , 10
210	plamo2 ,0 ,any ,RMS_NORM+MUL+ADD , 4
211	plamo2 ,0 ,any ,SSM_CONV+UNARY , 1
212	plamo3 ,0 ,any ,RMS_NORM+MUL , 9
213	plamo3 ,0 ,any ,RMS_NORM+MUL+ADD , 4
214	pockettts ,0 ,any ,NORM+MUL+ADD , 5
215	qwen ,0 ,any ,RMS_NORM+MUL , 5
216	qwen2 ,0 ,any ,RMS_NORM+MUL , 5
217	qwen2moe ,1 ,any ,ADD+ADD , 2
218	qwen2moe ,1 ,any ,MUL+ADD , 4
219	qwen2moe ,1 ,any ,RMS_NORM+MUL , 5
220	qwen2moe ,1 ,any ,SOFT_MAX+ARGSORT+GET_ROWS , 2
221	qwen2vl ,0 ,any ,RMS_NORM+MUL , 5
222	qwen3 ,0 ,any ,RMS_NORM+MUL , 9
223	qwen35 ,0 ,any ,GATED_DELTA_NET+CPY , 1
224	qwen35 ,0 ,any ,RMS_NORM+MUL , 8
225	qwen35 ,0 ,any ,RMS_NORM+SCALE , 2
226	qwen35 ,0 ,any ,SSM_CONV+UNARY , 1
227	qwen35moe ,1 ,any ,ADD+ADD , 2
228	qwen35moe ,1 ,any ,GATED_DELTA_NET+CPY , 1
229	qwen35moe ,1 ,any ,MUL+ADD , 4
230	qwen35moe ,1 ,any ,RMS_NORM+MUL , 8
231	qwen35moe ,1 ,any ,RMS_NORM+SCALE , 2
232	qwen35moe ,1 ,any ,SOFT_MAX+ARGSORT+GET_ROWS+SUM_ROWS+CLAMP+DIV, 2
233	qwen35moe ,1 ,any ,SSM_CONV+UNARY , 1
234	qwen3moe ,1 ,any ,MUL+ADD , 4
235	qwen3moe ,1 ,any ,RMS_NORM+MUL , 9
236	qwen3moe ,1 ,any ,SOFT_MAX+ARGSORT+GET_ROWS+SUM_ROWS+CLAMP+DIV, 2
237	qwen3next ,0 ,any ,ADD+ADD , 2
238	qwen3next ,0 ,any ,GATED_DELTA_NET+CPY , 1
239	qwen3next ,0 ,any ,MUL+ADD , 4
240	qwen3next ,0 ,any ,RMS_NORM+MUL , 8
241	qwen3next ,0 ,any ,RMS_NORM+SCALE , 2
242	qwen3next ,0 ,any ,SOFT_MAX+ARGSORT+GET_ROWS+SUM_ROWS+CLAMP+DIV, 2
243	qwen3next ,0 ,any ,SSM_CONV+UNARY , 1
244	qwen3tts ,0 ,any ,RMS_NORM+MUL , 9
245	qwen3vl ,0 ,any ,RMS_NORM+MUL , 9
246	qwen3vlmoe ,1 ,any ,MUL+ADD , 4
247	qwen3vlmoe ,1 ,any ,RMS_NORM+MUL , 9
248	qwen3vlmoe ,1 ,any ,SOFT_MAX+ARGSORT+GET_ROWS+SUM_ROWS+CLAMP+DIV, 2
249	qwen4exp ,0 ,any ,ADD+ADD+ADD , 1
250	qwen4exp ,0 ,any ,ADD+ADD+ADD+ADD+ADD+ADD+ADD , 9
251	qwen4exp ,0 ,any ,GATED_DELTA_NET+CPY , 1
252	qwen4exp ,0 ,any ,MUL+ADD , 4
253	qwen4exp ,0 ,any ,RMS_NORM+MUL , 13
254	qwen4exp ,0 ,any ,RMS_NORM+SCALE , 2
255	qwen4exp ,0 ,any ,SOFT_MAX+ARGSORT+GET_ROWS+SUM_ROWS+CLAMP+DIV, 2
256	qwen4exp ,0 ,any ,SSM_CONV+UNARY , 1
257	refact ,0 ,any ,RMS_NORM+MUL , 5
258	refact ,0 ,any ,RMS_NORM+MUL , 5
259	rnd1 ,0 ,any ,MUL+ADD , 4
260	rnd1 ,0 ,any ,RMS_NORM+MUL , 9
261	rnd1 ,0 ,any ,SOFT_MAX+ARGSORT+GET_ROWS+SUM_ROWS+CLAMP+DIV, 2
262	seed_oss ,0 ,any ,RMS_NORM+MUL , 5
263	smallthinker ,0 ,any ,MUL+ADD , 4
264	smallthinker ,0 ,any ,RMS_NORM+MUL , 5
265	smollm3 ,0 ,any ,RMS_NORM+MUL , 5
266	spark2_5 ,0 ,any ,RMS_NORM+MUL , 5
267	stablelm ,0 ,any ,NORM+MUL , 4
268	stablelm ,0 ,any ,NORM+MUL+ADD , 5
269	starcoder ,0 ,any ,NORM+MUL+ADD , 5
270	starcoder2 ,0 ,any ,NORM+MUL+ADD , 5
271	talkie ,0 ,any ,ADD+ADD , 2
272	xverse ,0 ,any ,RMS_NORM+MUL , 5

19 KiB Raw Permalink Blame History

19 KiB

Raw Permalink Blame History