Byte Magazine Unix benchmarking

General OpenPOWER Hardware > General CPU Discussion

<< < (2/2)

DKnoto:

--- Quote from: tle on June 07, 2024, 07:59:53 am ---
--- Quote from: DKnoto on June 06, 2024, 02:44:14 am ---My Talos II, Fedora 39:

--- Code: ---Architecture: ppc64le
Byte Order: Little Endian
CPU(s): 72
...
Socket(s): 1

--- End code ---

--- End quote ---
Could it be a Talos II Lite?

--- End quote ---

No, it's a full Talos II but with one CPU. When I feel that the hardware is already inefficient I will buy a second CPU ;)

So far the only program that is too slow on Talos is Firefox :(

helen:
Hi, this is very interesting!
Here's my Talos II Lite

--- Code: ---Architecture: ppc64le
Byte Order: Little Endian
CPU(s): 32
On-line CPU(s) list: 0-31
Model name: POWER9, altivec supported
Model: 2.3 (pvr 004e 1203)
Thread(s) per core: 4
Core(s) per socket: 8
Socket(s): 1
Frequency boost: enabled
CPU(s) scaling MHz: 100%
CPU max MHz: 3800.0000
CPU min MHz: 2166.0000
L1d cache: 256 KiB (8 instances)
L1i cache: 256 KiB (8 instances)
L2 cache: 4 MiB (8 instances)
L3 cache: 80 MiB (8 instances)
NUMA node(s): 1
NUMA node0 CPU(s): 0-31
Vulnerability Gather data sampling: Not affected
Vulnerability Itlb multihit: Not affected
Vulnerability L1tf: Mitigation; RFI Flush, L1D private per thread
Vulnerability Mds: Not affected
Vulnerability Meltdown: Mitigation; RFI Flush, L1D private per thread
Vulnerability Mmio stale data: Not affected
Vulnerability Reg file data sampling: Not affected
Vulnerability Retbleed: Not affected
Vulnerability Spec rstack overflow: Not affected
Vulnerability Spec store bypass: Mitigation; Kernel entry/exit barrier (eieio)
Vulnerability Spectre v1: Mitigation; __user pointer sanitization, ori31 speculation barrier enabled
Vulnerability Spectre v2: Mitigation; Software count cache flush (hardware accelerated), Software link stack flush
Vulnerability Srbds: Not affected
Vulnerability Tsx async abort: Not affected

--- End code ---

--- Code: --- BYTE UNIX Benchmarks (Version 5.1.3)

System: kvanneholtnipa: GNU/Linux
OS: GNU/Linux -- 6.9.3-gentoo -- #1 SMP Sun Jun 2 18:02:57 CEST 2024
Machine: ppc64le (PowerNV T2P9S01 REV 1.01)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
22:03:56 up 7 days, 3:53, 5 users, load average: 0.33, 0.13, 0.20; runlevel 2024-06-02

------------------------------------------------------------------------
Benchmark Run: Sun Jun 09 2024 22:03:56 - 22:32:05
32 CPUs in system; running 1 parallel copy of tests

Dhrystone 2 using register variables 42969120.6 lps (10.0 s, 7 samples)
Double-Precision Whetstone 4930.0 MWIPS (10.0 s, 7 samples)
Execl Throughput 1089.7 lps (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 55940.8 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 13843.0 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 203745.8 KBps (30.0 s, 2 samples)
Pipe Throughput 519706.9 lps (10.0 s, 7 samples)
Pipe-based Context Switching 56092.3 lps (10.0 s, 7 samples)
Process Creation 2319.9 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 2987.0 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 2006.8 lpm (60.0 s, 2 samples)
System Call Overhead 469894.7 lps (10.0 s, 7 samples)

System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 42969120.6 3682.0
Double-Precision Whetstone 55.0 4930.0 896.4
Execl Throughput 43.0 1089.7 253.4
File Copy 1024 bufsize 2000 maxblocks 3960.0 55940.8 141.3
File Copy 256 bufsize 500 maxblocks 1655.0 13843.0 83.6
File Copy 4096 bufsize 8000 maxblocks 5800.0 203745.8 351.3
Pipe Throughput 12440.0 519706.9 417.8
Pipe-based Context Switching 4000.0 56092.3 140.2
Process Creation 126.0 2319.9 184.1
Shell Scripts (1 concurrent) 42.4 2987.0 704.5
Shell Scripts (8 concurrent) 6.0 2006.8 3344.7
System Call Overhead 15000.0 469894.7 313.3
========
System Benchmarks Index Score 417.0

------------------------------------------------------------------------
Benchmark Run: Sun Jun 09 2024 22:32:05 - 23:00:27
32 CPUs in system; running 32 parallel copies of tests

Dhrystone 2 using register variables 424895900.7 lps (10.0 s, 7 samples)
Double-Precision Whetstone 118387.4 MWIPS (9.8 s, 7 samples)
Execl Throughput 11518.8 lps (29.9 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 564503.0 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 141555.5 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 2199353.4 KBps (30.0 s, 2 samples)
Pipe Throughput 5601329.9 lps (10.0 s, 7 samples)
Pipe-based Context Switching 978736.8 lps (10.0 s, 7 samples)
Process Creation 26301.2 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 23083.4 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 3091.5 lpm (60.2 s, 2 samples)
System Call Overhead 6228056.0 lps (10.0 s, 7 samples)

System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 424895900.7 36409.2
Double-Precision Whetstone 55.0 118387.4 21525.0
Execl Throughput 43.0 11518.8 2678.8
File Copy 1024 bufsize 2000 maxblocks 3960.0 564503.0 1425.5
File Copy 256 bufsize 500 maxblocks 1655.0 141555.5 855.3
File Copy 4096 bufsize 8000 maxblocks 5800.0 2199353.4 3792.0
Pipe Throughput 12440.0 5601329.9 4502.7
Pipe-based Context Switching 4000.0 978736.8 2446.8
Process Creation 126.0 26301.2 2087.4
Shell Scripts (1 concurrent) 42.4 23083.4 5444.2
Shell Scripts (8 concurrent) 6.0 3091.5 5152.5
System Call Overhead 15000.0 6228056.0 4152.0
========
System Benchmarks Index Score 4148.7

--- End code ---

I always felt my disks are very slow, and now I have proof. In single thread, they are about just 10% of the others here. Even with LUKS, they should not be THIS slow, should they?

Hasturtium:
That is a pretty stark difference. What’s your storage configuration?

helen:

--- Quote from: Hasturtium on June 09, 2024, 05:48:29 pm ---That is a pretty stark difference. What’s your storage configuration?

--- End quote ---

I have Btrfs on top of LUKS on top of Raw disk, no raid or stripe.

Hardware:
CPU --> PLX Technology PEX 9733 PCIe Gen 3 4x U.2 --> Western Digital Ultrastar DC SW630 NVMe SSD

I also have 2 other NVME-disks of different make and size, all are slow.

helen:

--- Quote from: helen on June 09, 2024, 05:10:51 pm ---Hi, this is very interesting!
Here's my Talos II Lite
- - 8< - - - - Cut

--- End quote ---

Seeing real numbers has finally gotten me motivated to clear out a disk, and test another distro.
Then I can also do some testing without full disk encryplion.
Will have to do some backup rotatoins first, thou :P

Navigation

[0] Message Index

[*] Previous page

Go to full version