Pulse report — Google Compute Engine n2¶

Submitted 2026-05-15T15:17:41+00:00 · fingerprint 33c40af047effdb0

Summary¶

26 ok · 2 skipped · 1 failed

benchmark	status	duration	summary
`cpu_python`	ok	0.45s	sieve 76.90 Mops/s, arith 11.83 Mops/s
`cpu_hash`	ok	4.01s	SHA-256 single-thread (16 MiB chunks): 1365 MiB/s
`cpu_compress`	ok	13.45s	zlib c/d 33/207 MiB/s, lzma 3 MiB/s, zlib ratio 0.66
`sqlite_oltp`	ok	8.01s	WAL mode: inserts=43522 selects=2851 updates=19208 ops/s
`memory`	ok	2.11s	memcpy 11527 MiB/s (ctypes), 1MiB alloc 48358/s
`disk_seq`	failed	0.14s	OSError(22, 'Invalid argument')
`disk_rand`	ok	19.59s	random 4K read: 1021 IOPS (O_DIRECT)
`network_download`	ok	0.54s	max 58.3 MiB/s (489 Mbps), median 55.7 MiB/s, σ=6.3 (3 samples)
`llm_inference`	ok	5.84s	tinyllama-1.1b-chat-v1.0.Q4_K_M.gguf on llama-cpp-python (32 threads): pp512=266.7 tok/s · tg128=44.4 tok/s
`gpu_compute`	ok	0.23s	CPU (scipy-openblas) (cpu-numpy): 755.0 GFLOPS @ 2048^3
`sysbench_cpu`	ok	30.04s	sysbench cpu @ 32t, prime=20000: 23861 events/s
`sysbench_memory`	ok	0.12s	sysbench memory read: 39322 MiB/s
`sysbench_threads`	ok	20.03s	sysbench threads @ 16t: 3177 events/s
`stream`	ok	1.14s	STREAM Triad 54.4 GB/s (Copy 47.7 · Scale 50.7 · Add 57.9) · 1144 MiB working set
`fio_disk`	ok	63.51s	4K randread 38829 IOPS · 1M seqwrite 733 MiB/s
`stress_ng_cpu`	ok	30.11s	stress-ng cpu @ 32t (matrixprod): 14416 bogo-ops/s
`stress_ng_memory`	ok	30.10s	stress-ng memrate: 2 bogo-ops/s
`7zip`	ok	38.94s	compress 100 MIPS, decompress 3812 MIPS
`openssl_speed`	ok	18.02s	AES-256-GCM (16K blocks): 3407 MiB/s
`c_ray`	ok	9.00s	c-ray-mt 7680x4320 rays=16 threads=32: 8.20 s
`ffmpeg_x264_encode`	ok	5.97s	x264 encode 900 frames @ 1080p in 6.0s = 150.7 fps (last reported: 153.0 fps)
`blender_bmw27`	ok	127.18s	BMW27 (CYCLES, 100% resolution): 125.6 s (wall: 127.2 s)
`glmark2`	skipped	0.00s	no DISPLAY/WAYLAND_DISPLAY — needs a graphical session
`docker`	ok	4.55s	pull 1.90s · run start (median) 529 ms
`compile_sqlite`	ok	29.54s	sqlite3.c compiled in 29.54s (8661 lines/s) using cc
`compile_cpp`	ok	2.98s	json.hpp compiled in 2.98s (8305 header lines/s) using c++
`compile_java`	ok	25.03s	javac compiled 500 classes / 786000 LOC in 25.03s (31400 lines/s)
`compile_redis`	ok	13.30s	redis built in 13.3s with -j8 (151 object files)
`browser_js`	skipped	0.00s	no chromium-family browser found

Subscores¶

Geomean of the per-bench headline metric within each category, normalised against the reference baseline (100 = parity).

category	score
compress	0.1
cpu	31.2
crypto	42.6
disk	0.5
gpu	6.3
media	75.4
memory	161.5
ml	55.5
render	68.2

System¶

field	value
platform	`linux/ubuntu/x86_64`
CPU	AMD EPYC 7B13 (32 threads)
RAM	125.8 GiB
motherboard	Google Google Compute Engine
NUMA	1 node(s)
Submitter	Patrik maly.patrik007@gmail.com

Detail¶

`cpu_python` — cpu¶

status: ok · duration: 0.45s

sieve 76.90 Mops/s, arith 11.83 Mops/s

metric	value	unit
sieve_throughput	76.9	Mops/s
arith_throughput	11.83	Mops/s
primes_found	1.489e+05	primes

`cpu_hash` — cpu¶

status: ok · duration: 4.01s

SHA-256 single-thread (16 MiB chunks): 1365 MiB/s

metric	value	unit
sha256_throughput	1365	MiB/s
chunk_size_mib	16	MiB

`cpu_compress` — cpu¶

status: ok · duration: 13.45s

zlib c/d 33/207 MiB/s, lzma 3 MiB/s, zlib ratio 0.66

metric	value	unit
zlib_compress	32.71	MiB/s
zlib_decompress	207.1	MiB/s
lzma_compress	2.598	MiB/s
zlib_ratio	0.665	ratio

`sqlite_oltp` — database¶

status: ok · duration: 8.01s

WAL mode: inserts=43522 selects=2851 updates=19208 ops/s

metric	value	unit
inserts_per_s	4.352e+04	ops/s
selects_per_s	2851	ops/s
updates_per_s	1.921e+04	ops/s
txn_per_s	6.558e+04	ops/s

`memory` — memory¶

status: ok · duration: 2.11s

memcpy 11527 MiB/s (ctypes), 1MiB alloc 48358/s

metric	value	unit
memcpy	1.153e+04	MiB/s
alloc_1mib_per_sec	4.836e+04	allocs/s

`disk_seq` — disk¶

status: failed · duration: 0.14s

`disk_rand` — disk¶

status: ok · duration: 19.59s

random 4K read: 1021 IOPS (O_DIRECT)

metric	value	unit
rand_read_iops	1021	IOPS
rand_read_throughput	3.988	MiB/s
file_size_mib	512	MiB

`network_download` — network¶

status: ok · duration: 0.54s

max 58.3 MiB/s (489 Mbps), median 55.7 MiB/s, σ=6.3 (3 samples)

metric	value	unit
download_max_mibps	58.3	MiB/s
download_median_mibps	55.71	MiB/s
download_max_mbps	489	Mbps
download_median_mbps	467.3	Mbps
download_stdev_mibps	6.269	MiB/s
download_samples	3	n

`llm_inference` — ml¶

status: ok · duration: 5.84s

tinyllama-1.1b-chat-v1.0.Q4_K_M.gguf on llama-cpp-python (32 threads): pp512=266.7 tok/s · tg128=44.4 tok/s

metric	value	unit
llm_pp512_tokens_per_sec	266.7	tok/s
llm_tg128_tokens_per_sec	44.41	tok/s

`gpu_compute` — gpu¶

status: ok · duration: 0.23s

CPU (scipy-openblas) (cpu-numpy): 755.0 GFLOPS @ 2048^3

metric	value	unit
gpu_sgemm_gflops	755	GFLOPS
gpu_matmul_size	2048	N

`sysbench_cpu` — cpu¶

status: ok · duration: 30.04s

sysbench cpu @ 32t, prime=20000: 23861 events/s

metric	value	unit
sysbench_cpu_events_per_sec	2.386e+04	events/s
sysbench_cpu_threads	32	threads

`sysbench_memory` — memory¶

status: ok · duration: 0.12s

sysbench memory read: 39322 MiB/s

metric	value	unit
sysbench_memory_read	3.932e+04	MiB/s

`sysbench_threads` — cpu¶

status: ok · duration: 20.03s

sysbench threads @ 16t: 3177 events/s

metric	value	unit
sysbench_threads_events_per_sec	3177	events/s
sysbench_threads_count	16	threads

`stream` — memory¶

status: ok · duration: 1.14s

STREAM Triad 54.4 GB/s (Copy 47.7 · Scale 50.7 · Add 57.9) · 1144 MiB working set

metric	value	unit
stream_copy_gb_s	47.72	GB/s
stream_scale_gb_s	50.72	GB/s
stream_add_gb_s	57.89	GB/s
stream_triad_gb_s	54.41	GB/s

`fio_disk` — disk¶

status: ok · duration: 63.51s

4K randread 38829 IOPS · 1M seqwrite 733 MiB/s

metric	value	unit
fio_randread_iops	3.883e+04	IOPS
fio_randread_throughput	151.7	MiB/s
fio_seqwrite_throughput	732.5	MiB/s

`stress_ng_cpu` — cpu¶

status: ok · duration: 30.11s

stress-ng cpu @ 32t (matrixprod): 14416 bogo-ops/s

metric	value	unit
stress_ng_cpu_bogo_ops_per_sec	1.442e+04	bogo-ops/s
stress_ng_cpu_threads	32	threads

`stress_ng_memory` — memory¶

status: ok · duration: 30.10s

stress-ng memrate: 2 bogo-ops/s

metric	value	unit
stress_ng_memrate_bogo_ops_per_sec	1.73	bogo-ops/s

`7zip` — compress¶

status: ok · duration: 38.94s

compress 100 MIPS, decompress 3812 MIPS

metric	value	unit
7z_compress_rating	100	MIPS
7z_decompress_rating	3812	MIPS

`openssl_speed` — crypto¶

status: ok · duration: 18.02s

AES-256-GCM (16K blocks): 3407 MiB/s

metric	value	unit
aes256_gcm	3407	MiB/s

`c_ray` — render¶

status: ok · duration: 9.00s

c-ray-mt 7680x4320 rays=16 threads=32: 8.20 s

metric	value	unit
c_ray_render_time_s	8.202	s
c_ray_threads	32	threads

`ffmpeg_x264_encode` — media¶

status: ok · duration: 5.97s

x264 encode 900 frames @ 1080p in 6.0s = 150.7 fps (last reported: 153.0 fps)

metric	value	unit
ffmpeg_x264_encode_fps	150.7	fps
ffmpeg_x264_output_mib	22.9	MiB

`blender_bmw27` — render¶

status: ok · duration: 127.18s

BMW27 (CYCLES, 100% resolution): 125.6 s (wall: 127.2 s)

metric	value	unit
blender_bmw27_render_seconds	125.6	s

`glmark2` — gpu¶

status: skipped · duration: 0.00s

skipped: no DISPLAY/WAYLAND_DISPLAY — needs a graphical session

`docker` — container¶

status: ok · duration: 4.55s

pull 1.90s · run start (median) 529 ms

metric	value	unit
docker_pull_time	1.897	s
docker_run_median	0.5293	s
docker_run_min	0.5169	s

`compile_sqlite` — compile¶

status: ok · duration: 29.54s

sqlite3.c compiled in 29.54s (8661 lines/s) using cc

metric	value	unit
compile_time	29.54	s
compile_throughput	8661	lines/s
download_time	0.7005	s

`compile_cpp` — compile¶

status: ok · duration: 2.98s

json.hpp compiled in 2.98s (8305 header lines/s) using c++

metric	value	unit
cpp_compile_time	2.982	s
cpp_compile_throughput	8305	lines/s
download_time	0.4435	s

`compile_java` — compile¶

status: ok · duration: 25.03s

javac compiled 500 classes / 786000 LOC in 25.03s (31400 lines/s)

metric	value	unit
javac_time	25.03	s
javac_lines_per_sec	3.14e+04	lines/s
javac_classes_per_sec	19.97	classes/s

`compile_redis` — compile¶

status: ok · duration: 13.30s

redis built in 13.3s with -j8 (151 object files)

metric	value	unit
redis_build_time	13.3	s
redis_objs_per_sec	11.36	objs/s
redis_jobs	8	threads
download_time	0.2054	s

`browser_js` — browser¶

status: skipped · duration: 0.00s

skipped: no chromium-family browser found

Pulse report — Google Compute Engine n2¶

Summary¶

Subscores¶

System¶

Detail¶

cpu_python — cpu¶

cpu_hash — cpu¶

cpu_compress — cpu¶

sqlite_oltp — database¶

memory — memory¶

disk_seq — disk¶

disk_rand — disk¶

network_download — network¶

llm_inference — ml¶

gpu_compute — gpu¶

sysbench_cpu — cpu¶

sysbench_memory — memory¶

sysbench_threads — cpu¶

stream — memory¶

fio_disk — disk¶

stress_ng_cpu — cpu¶

stress_ng_memory — memory¶

7zip — compress¶

openssl_speed — crypto¶

c_ray — render¶

ffmpeg_x264_encode — media¶

blender_bmw27 — render¶

glmark2 — gpu¶

docker — container¶

compile_sqlite — compile¶

compile_cpp — compile¶

compile_java — compile¶

compile_redis — compile¶

browser_js — browser¶