RandomHash2 CPU benchmark: post your hashrate (results table inside)

[parniancoin]

RandomHash2 CPU benchmark thread

RandomHash2 is the CPU-oriented proof-of-work used by PascalCoin v5 and ParnianCoin (PARC). It chains randomly selected hash algorithms with memory-dependent steps, so performance depends a lot on cache size, memory speed and core count, not just clock speed.

There is very little public data on how different CPUs actually perform on RandomHash2. This thread collects real benchmark results so anyone can compare CPUs.

How to run the benchmark (about 20 seconds, no account or pool needed)

1. Download parnianminer (open source, GPLv3, no dev fee):
github.com/ParnianCoin/parniancoin-miner/releases
Verify the SHA-256 checksum on the release page.

2. Run:

Code:

parnianminer -cpu -cputhreads 0 -testperformance 20

(-cputhreads 0 = auto. Run parnianminer -completelist to see your logical cores.)

3. Post your result using this template:

Code:

CPU:        (e.g. Ryzen 7 5800X)
Threads:    (e.g. 16)
OS:         (Windows / Linux + distro)
Hashrate:   (H/s or kH/s)
Power:      (watts, optional)
Notes:      (RAM speed, undervolt, etc. optional)

Results table
I will keep a sorted table of all submitted results in this first post.

CPU	Threads	OS	Hashrate	Watts
(first results coming)

Questions about RandomHash2 tuning (thread count, Linux vs Windows, the ~7% faster "modern" Linux build) are welcome here.

[parniancoin]

First result to get the table started, from my own laptop:

Code:

CPU:        Intel Core i7-4710HQ @ 2.50GHz (Haswell, 4 cores / 8 threads)
Threads:    8 (auto)
OS:         Windows 10/11 x64
Hashrate:   106,119 H/s (~106 kH/s)
Power:      not measured (47W TDP laptop chip)
Notes:      parnianminer 2.4, -testperformance 20, default settings

Not bad for a 2014 laptop CPU. RandomHash2 seems to like Haswell's cache layout more than I expected.

Curious how modern Ryzen and Intel chips compare. If you run the test, please also try fewer threads (e.g. half your logical cores): on some CPUs RandomHash2 scales better on physical cores than on hyperthreads.