AstroBWT Benchmarks

[DankFC]

Please leave your CPU benchmarks with OS, miner, and hashrate here.

Ex.

AMD Ryzen Threadripper
OS: Linux, Ubuntu
Miner: Official v2
Hashrate: 560 H/s

[Peppinux]

CPU: Intel Core i7-6700 @ 3.40GHz (4 cores/8 threads)
RAM: 16GB DDR4 CL14 @ 2133MHz

Official Miner v1
OS: Ubuntu 19 x64
Hashrate: ~105 H/s

Official Miner v2 (Optimized)

OS: Ubuntu 19 x64
Hashrate: ~375 H/s

XMRig v5.8.1 (gcc/9.2.0)
OS: Windows 10 x64
Hashrate: ~335 H/s

XMRig v5.8.2 & XMRig v5.9.0 (gcc/9.2.0)

OS: Windows 10 x64
Hasharate: ~420 H/s

[vladimwd]

Intel® Xeon® E5-4657L v2
127-160h/s -12/24 ubuntu 17.04

[pi3rk]

Intel(R) Core(TM) i7-87s50H CPU @ 2.20GHz
6 cores/12 threads
~75H/s

[andy1999ch]

AMD RYZEN 7 3700 X

200-230 h/s

[dillonyap]

i5-9300H . 70H ~ 95H

[Lebowski1234]

AMD Ryzen 5 3600 (3.6GHz) ~180 H/S
Intel Xeon E31240 (3.3GHz) ~60 H/S
Intel Core i5-2500K (3.3GHz) ~50 H/S
AMD A4-7300 (3.8GHz) ~12 H/S

[torrantz]

ryzen 5 3600

ryzen @ stock 180-200 H/s 26 watt
ryzen @ 3950 220-240 H/s 34 watt
ryzen @ 4150 230-250 H/s 48 watt

[Peterlustig42]

Ryzen 5 2600 @ 3900 140-160 H/s
Ryzen 3 1200 @ 2000 45-55 H/s
Ryzen 3 120 @ 3900 80-90 H/s

[Notoriousjoshyb]

I7-7700k 100H/S - 120H/S

[JaernStaang]

Intel Core i7-6700K @4.20GHz 120-138H/s
Intel Atom x5-Z8330 @ 1.44Ghz between 2-6H/s
Intel Core M 5Y10 @ 0.80Ghz between 25-33H/s

[jackcoble]

i7-7700HQ @ 2.80GHz
~ 90H/s

[ghost]

i3 8100(4 core),using 3 core with proxmox virtualization,
~75H/s

[gfox72]

Intel i5-4570 @ 3.2GhZ (iMac late 2013 8Gb)
~56.5H/s

[Nelbert442]

Intel i7-3770K @3.5GHz
~72 - ~92H/s averaging around about 82H/s

[gfox72]

Intel i3-550 @3.2GHz (4Gb)
~12.8H/s

[gfox72]

Intel i7-9400K @ 3.6GHz (16Gb)
~109.2H/s
(did have fortnite run in the background)

[Filter-999]

Ryzen 2700 3.9GHz
~172 H/s

[DankFC]

AMD Ryzen Threadripper 1950X (Linux)
32 Threads: ~129 H/s
15 threads: ~240 H/s
6 Threads: ~142 H/s

[phucpn]

Intel Xeon E5 2660 v4 20/28 threads
200 H/s

[criptocapitan]

Intel Core i5-3210M CPU 2.5 GHz ~35 H/s

[SChernykh]

Ryzen 7 3700X, 4.1 GHz 315 H/s
https://i.imgur.com/G22iHhW.png

[jrb0043]

Ryzen 7 3800X 4.4Ghz 970H/s

https://i.imgur.com/ND4PI1e.png

[PikkonMG]

Qualcomm SDM845 Snapdragon 845 (10 nm) CPU Octa-core (4x2.8 GHz Kryo 385 Gold & 4x1.7 GHz Kryo 385 Silver)
Ubuntu XMRigCC 220 H/s

Ryzen 7 2700x DDR4 2666 "Stock settings"
Windows 10 XMRig-5.9.0 610 H/s

[DankFC]

AMD Ryzen Threadripper 1950X
OS: Linux
Miner: Official v2
Hashrate: 560 H/s

[Nelbert442]

Updated for latest dero-miner version:

i7-3770k @3.5GHz (non-oc), Windows 10

~240 - 245H/s <--- dero-miner (latest binaries)
~242H/s <--- xmrig (v5.8.1)
~292 - 299H/s <--- xmrig (v5.8.2)

[revelacaogr]

i9 9900k@5.2Ghz
Windows 10-7-ubuntu mint
dero-miner (latest binaries)- XMRig-5.9.0 610 H/s

i7 4790k@4.40 GHz
Windows 10-7-ubuntu mint
dero-miner (latest binaries)- XMRig-5.9.0 300 H/s

[ghost]

AMD RYZEN 7 1700@3.7Ghz
Windows 10
XMRig-5.9.0

570 h/s

[Atceton]

AMD FX-4150 95W OverClock 4300 MHz, 4 GB RAM DDR3-1333 - 134 h/s
Windows 10 x64 LTSC build 1809, disabled telemetry.

Intel Core i3 3120M 2500 MHz 8 GB RAM DDR3-1600L - 136 h/s
Windows 10 x64 LTSB build 1607, disabled telemetry.

All XMRig-5.9.0.

[MonstaCoin]

Ryzen 9 3950x , 2x 8gb DDR4 Predators 3200mhz, Coolmaster water cooler and 240mm radiator. I got about 860H/s

[Shekelme]

AMD FX-8350, Win7 x64 Ultimate, 24 Gb RAM, xmrig-5.11.2, 8 threads - 223 H/s

[netmonk]

model name : Intel(R) Xeon(R) CPU E5-2620 v4 @ 2.10GHz
[2020-11-28 09:16:54.280] speed 10s/60s/15m 679.0 677.8 679.9 H/s max 710.6 H/s

[Notoriousjoshyb]

RTX 3090 FE - XMRIG speed 10s/60s/15m 6632.6 6553.5 n/a H/s max 6825 H/s
AMD Ryzen 9 3900x - XMRIG speed 10s/60s/15m 1127.0 n/a n/a H/s max 1149.4 H/s

[Afkbio]

Intel 9900K 670H/s

[DJrunkie]

1080ti 1949mhz cpu 5599mhz memory on Windows 10 XMRIG 6.7.2 10s/60s/15m 2517.8 2585.3 2568.1 max 2697 H/s

[DJrunkie]

Ryzen 9 3900x XMRIG 6.7.2 on Windows 10 with HUGE PAGES Permission granted & Enabled 1381 H/s max

And here is the config I'm using.

"cpu": {
    "enabled": true,
    "huge-pages": true,
    "huge-pages-jit": true,
    "hw-aes": null,
    "priority": null,
    "memory-pool": false,
    "yield": true,
    "asm": true,
    "argon2-impl": null,
    "astrobwt-max-size": 400,
    "astrobwt-avx2": true,
    "argon2": [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23],
    "astrobwt": [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23],

Huge pages adds about 5%
Huge pages jit adds about .3%
astrobwt-avx2 true adds about 5-10%
astrobwt-max-size 400 adds about 20%

I noticed the most important setting is astrobwt-max-size. 400 or 410 works best on this ryzen. 560 works best on a Xeon E5-1650, and 580 works best on a I5-4670k. Using the default of 550 on the Ryzen results in about 20% lower scores or about 1123 H/s Max. Before all the other adjustments I was getting about 950-1000 H/s

--astrobwt-max-size=N | skip hashes with large stage 2 size, default: 550, min: 400, max: 1200

If you adjust the astrobwt-max-size you must stop and start xmrig. It says its reloading the config file when you change it but the hash rate will be unchanged. Also with astrobwt-max-size set incorrectly I could get as low as 700 H/s so it does take some fine tuning, no idea on how to determine the best other than trial and error. Because of this I recommend changing this setting in the config from 60 seconds to 20 or 10 seconds.

"print-time": 20,

Cheers!

Update October 5th 2021: With seeing Deros success I have revisited this and made some major improvements to the hashrate on my Ryzen 3900x. I started disabling threads because I just wanted some idle CPU power while working. However I noticed my hashrate actually started going up and not down when disabling threads. The max hashrate I could achieve was 1782.5 H/s with 19 threads enabled. However I also got 1778.9 H/s with 18 threads enabled, which I think realistically is the most you want to use.

from 24 threads :
"astrobwt": [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23],

to 18 threads:
"astrobwt": [0, 2, 3, 4, 6, 7, 8, 9, 11, 12, 13, 15, 16, 17, 18, 20, 21, 23],

I'm guessing too many cores enabled is producing too much heat, so you're actually hurting performance. People with the Ryzen 7 and Threadrippers should try this to see if they get an improvement from having less cores enabled. Keep in mind I had at least half the core enabled so it was using 50% minimum for each core.

I felt that 18 threads was still too much while I'm working. Now I have 16 active threads and get 1703.4 H/s max
"astrobwt": [0, 2, 3, 4, 6, 8, 10, 11, 12, 13, 15, 16, 18, 20, 21, 23],

However 15 threads may be more efficient with 1627.7H/s max
"astrobwt": [0, 2, 3, 4, 6, 8, 10, 12, 13, 15, 16, 18, 20, 21, 23],

So if you have tons of threads try disabling some of the cores until you stop seeing hash improvements.

[bezatine]

Intel Celeron N3060 2 cores 2 threads

56.67 max. 61.79 H/s

Feb 22 2021 watch the fork difficulty that happened I think it may have opened things up for older processors, literally wanna go buy 10 laptops and I'll equal a ryzen right lmao,

[netmonk]

Ryzen 9 3900x XMRIG 6.7.2 on Windows 10 with HUGE PAGES Permission granted & Enabled 1381 H/s max

And here is the config I'm using.

"cpu": {
    "enabled": true,
    "huge-pages": true,
    "huge-pages-jit": true,
    "hw-aes": null,
    "priority": null,
    "memory-pool": false,
    "yield": true,
    "asm": true,
    "argon2-impl": null,
    "astrobwt-max-size": 400,
    "astrobwt-avx2": true,
    "argon2": [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23],
    "astrobwt": [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23],

Huge pages adds about 5%
Huge pages jit adds about .3%
astrobwt-avx2 true adds about 5-10%
astrobwt-max-size 400 adds about 20%

I noticed the most important setting is astrobwt-max-size. 400 or 410 works best on this ryzen. 560 works best on a Xeon E5-1650, and 580 works best on a I5-4670k. Using the default of 550 on the Ryzen results in about 20% lower scores or about 1123 H/s Max. Before all the other adjustments I was getting about 950-1000 H/s

--astrobwt-max-size=N | skip hashes with large stage 2 size, default: 550, min: 400, max: 1200

If you adjust the astrobwt-max-size you must stop and start xmrig. It says its reloading the config file when you change it but the hash rate will be unchanged. Also with astrobwt-max-size set incorrectly I could get as low as 700 H/s so it does take some fine tuning, no idea on how to determine the best other than trial and error. Because of this I recommend changing this setting in the config from 60 seconds to 20 or 10 seconds.

"print-time": 20,

Cheers!

Using this post and with trial and error i reached
[2021-02-24 08:16:29.520] miner speed 10s/60s/15m 587.4 590.5 590.8 H/s max 607.3 H/s

Using
sudo ./xmrig -p w=Rig2 -a astrobwt -k --astrobwt-max-size=550 --astrobwt-avx2 --asm=ryzen

On AMD 1950x threadripper, by default i get 555 to 560 on average.

[StormGoth]

I7 6600u 32GB, Xmrig 6.15.2-mo2, Linux, compiled with GCC 11.2.1, O3 and pgo, lto, astrobwt-avx2, max-size 580, huge pages, 1GB pages: 225-230 H/s.

With a generic build/config i got about 170, so not too shabby...

[StormGoth]

Ryzen 7, 5800u, 32 GB, same setup as the 6600u, 720 H/s

[Gtreii]

Corei5-2400. 223h

[thespicer21]

Still tweaking but so far :

Ryzen 5 3600 (8GB / 11 Threads / avx2 true / max-size 550 / Huge Pages off) = 1178 h/s
Ryzen 5 3600 (8GB / 11 Threads / avx2 true / max-size 550 / Huge Pages off) = 1135 h/s
i7 4790 (16GB / 7 Threads / avx2 true / max-size 570 / Huge Pages on) = 453 h/s
i7 4790 (8GB / 7 Threads / avx2 true / max-size 570 / Huge Pages off) = 442 h/s
Xeon E3 1245 v3 (8GB / 7 Threads / avx2 true / max-size 550 / Huge Pages on) = 434 h/s
i5 4590 (16GB / 4 Threads / avx2 true / max-size 550 / Huge Pages on) = 315 h/s
i3 4160 (8GB / 4 Threads / avx2 true / max-size 550 / Huge Pages off) = 235 h/s

In my experience the amount of RAM you have is pretty much irrelevant as you can see from the figures above.

Same goes for Huge Pages. Most of my rigs have 60GB SSDs so can't enable Huge Pages and it doesn't seem to make any difference at all compared to the rigs that have bigger drives with HP enabled.

By far the biggest improvement I've found for every processor is by switching avx2 to true - but it will raise your core temps.

Also in most cases you want to run at one thread less than your maximum capability. It will either improve your hash rate slightly or give you the same hash rate for less power while still allowing your CPU to do other stuff (in my case the other stuff is runing gMiner with GPUs mining Flux)

Max-size can add a little bit of juice too but you have to play around with it, one step in the wrong direction will lower your hashrate.

Also worth noting the difference between the two R5 3600 rigs. They are identical hardware and config wise but one processor gets 40h/s more than the other. I guess that's just the CPU lottery though ;)

[HoshiyominoUC]

Mediatek Helio P20 MT6757 Termux clang 53.1 H/s
Qualcomm S4 Pro APQ8064 Bus error
I don't know why it returns Bus error on ARMv7 but running well in ARMv8+
It happened during benchmark of xmrig-v6.16.2-mo2

[Megaminer2018]

3900x 6.5 kh
5900x 6.9 kh

[kox55rus]

7950x curva -0.1v boot 4700 all core - 90-92watt - 20.5 kh/s

[NVMDSTEVil]

2990wx @ 3.6ghz 22.2KH
3900x @ 4.2ghz 11.7KH
5900x @ 4.6ghz 14KH

Uranus 003 miner, windows