RandomX Benchmarks! https://monerobenchmarks.info - page 8.

seefatlow

jr. member

Activity: 80

Merit: 1

Quote from: joblo on November 03, 2019, 07:46:04 PM

Quote from: seefatlow on November 03, 2019, 03:59:29 AM

The issue here is the amount of hardware that needs to be bought for a CPU algo. For a GPU algo, all you need to increase hashrates is to buy the GPU and riser card. Now, you will have to get the whole shebang (PSU, memory, MB etc.) How will this be ever profitable? Will there ever be a riser card/board equivalent for CPUs?

That's all part of the cost per hash. It's no different than an algo with a lower hashrate requiring
more CPU (or GPU) power to produce the same hash. What matters is how much the hash is worth.

Yes Intel has Xeon Phi compute addin boards with up to 72 cores with 4:1 hyperthreading
for 288 threads. The're compute beasts but very expensive and not suitable for mining due
to the low mem/thread ratio. It maxes out at 384 GB which is similar to the cache/thread
ratio of most desktop CPUs.

If Loki network hashrate is any indication, it's not going to be profitable for RandomX

joblo

legendary

Activity: 1470

Merit: 1114

Quote from: seefatlow on November 03, 2019, 03:59:29 AM

The issue here is the amount of hardware that needs to be bought for a CPU algo. For a GPU algo, all you need to increase hashrates is to buy the GPU and riser card. Now, you will have to get the whole shebang (PSU, memory, MB etc.) How will this be ever profitable? Will there ever be a riser card/board equivalent for CPUs?

That's all part of the cost per hash. It's no different than an algo with a lower hashrate requiring
more CPU (or GPU) power to produce the same hash. What matters is how much the hash is worth.

Yes Intel has Xeon Phi compute addin boards with up to 72 cores with 4:1 hyperthreading
for 288 threads. The're compute beasts but very expensive and not suitable for mining due
to the low mem/thread ratio. It maxes out at 384 GB which is similar to the cache/thread
ratio of most desktop CPUs.

herrdrone

sr. member

Activity: 633

Merit: 267

Quote from: badbart on November 03, 2019, 11:24:19 AM

Any idea on profits? I set up one randomX rig and its mining loki, making .47 cents a day, I hope xmr is more.

Profitability on RandomX, only after fork is done on November 30 so you will have: network hashrate and monero price to do the math.

sundownz06

newbie

Activity: 110

Merit: 0

Shoot... It may be time to fire up my 100+ Xeon X5660 CPUs and my bank of Dell R815 servers again (120+ Opteron 6380).

Exciting stuff... I've had a lot of my farm turned off lately.

badbart

member

Activity: 449

Merit: 24

Any idea on profits? I set up one randomX rig and its mining loki, making .47 cents a day, I hope xmr is more.

seefatlow

jr. member

Activity: 80

Merit: 1

Quote from: Marvell2 on November 02, 2019, 02:13:52 PM

Quote from: seefatlow on November 02, 2019, 12:00:44 AM

Makes no sense to me why anyone should be excited about coins moving to RandomX. The cost per added hashrate for RandomX is so much higher than that for a GPU mineable coin and if you think the network hashrate of a RandomX coin will be low, think again. Just take a look at the Loki network hashrate as an example of what's in store for all CPU mineable coins. It's 71 MH vs 11KH from a single Ryzen 3900X!! How many Ryzen 3900X do you need to make enough coins to ROI?

not sure what your issue is here, yeah the ryzen is a beast at random x , but its a beast at every other cpu algo out there anyways. its 12 cores on a consumer chip basically we have never seen this type of power for non server builds.

i mean its expected it would destroy every cpu tuned algo out there, which random x is and it does, you can buy a
ryzen 3600 for 229 and get 6k on randomx , not sure what the issue is

The issue here is the amount of hardware that needs to be bought for a CPU algo. For a GPU algo, all you need to increase hashrates is to buy the GPU and riser card. Now, you will have to get the whole shebang (PSU, memory, MB etc.) How will this be ever profitable? Will there ever be a riser card/board equivalent for CPUs?

Marvell2

full member

Activity: 1148

Merit: 132

Quote from: seefatlow on November 02, 2019, 12:00:44 AM

Makes no sense to me why anyone should be excited about coins moving to RandomX. The cost per added hashrate for RandomX is so much higher than that for a GPU mineable coin and if you think the network hashrate of a RandomX coin will be low, think again. Just take a look at the Loki network hashrate as an example of what's in store for all CPU mineable coins. It's 71 MH vs 11KH from a single Ryzen 3900X!! How many Ryzen 3900X do you need to make enough coins to ROI?

not sure what your issue is here, yeah the ryzen is a beast at random x , but its a beast at every other cpu algo out there anyways. its 12 cores on a consumer chip basically we have never seen this type of power for non server builds.

i mean its expected it would destroy every cpu tuned algo out there, which random x is and it does, you can buy a
ryzen 3600 for 229 and get 6k on randomx , not sure what the issue is

herrdrone

sr. member

Activity: 633

Merit: 267

Quote from: seefatlow on November 02, 2019, 12:00:44 AM

Makes no sense to me why anyone should be excited about coins moving to RandomX. The cost per added hashrate for RandomX is so much higher than that for a GPU mineable coin and if you think the network hashrate of a RandomX coin will be low, think again. Just take a look at the Loki network hashrate as an example of what's in store for all CPU mineable coins. It's 71 MH vs 11KH from a single Ryzen 3900X!! How many Ryzen 3900X do you need to make enough coins to ROI?

Guess you are late to discussion, as RandomX as well as others options were discussed to exhaustion months ago:

https://github.com/monero-project/meta/issues/316#issuecomment-472677985
https://github.com/monero-project/meta/issues/321

Unlike zcash, eth & others that have done absolutely nothing when their networks got overtaken by ASICs.

Monero is moving forward by implementing RandomX, properly tested and audited (https://www.reddit.com/r/Monero/comments/ckhuq9/randomx_audit_status_final/).

seefatlow

jr. member

Activity: 80

Merit: 1

Makes no sense to me why anyone should be excited about coins moving to RandomX. The cost per added hashrate for RandomX is so much higher than that for a GPU mineable coin and if you think the network hashrate of a RandomX coin will be low, think again. Just take a look at the Loki network hashrate as an example of what's in store for all CPU mineable coins. It's 71 MH vs 11KH from a single Ryzen 3900X!! How many Ryzen 3900X do you need to make enough coins to ROI?

herrdrone

sr. member

Activity: 633

Merit: 267

RandomX Testnet pools

Xmrig
randomx-benchmark.xmrig.com:7777
Github: https://github.com/xmrig/xmrig/issues/1204

supportXMR
69.164.198.226:3333
Site: http://69.164.198.226/
Related Post:https://www.reddit.com/r/MoneroMining/comments/do6hyp/supportxmr_randomx_testnet_pool/

minexmr.com
rx.minexmr.com:4444
Site: https://rx.minexmr.com/
Related Post: https://www.reddit.com/r/MoneroMining/comments/dfxpjy/minexmr_randomx_testnet_pool/

Ursul0

sr. member

Activity: 857

Merit: 262

Quote from: WhyMe on October 24, 2019, 10:36:39 AM

Quote from: Ursul0 on October 24, 2019, 07:36:16 AM

2*Xeon E5 2660 v2 give 7330h/s on RandomX with 1 thread per core (L3 cache = 25MB*2slots)
setting higher thread count will degrade performance of paired(on a single core) threads

2* Xeon E5645 ( total is 12c/24t @2.4GHz )
Default setting is 12t and give 2220h/s. Setting other than 12t gives lower hashrate.
6t = 1680h/s
8t = 2030h/s
16t = 1920h/s
20t = 1640h/s ( and with 20t server is already 98-99% cpu usage )

I have 10 cores, 20 threads on each E5 and 25MB L3 cache
randomx needs 2MB scratchpad (cache) and 2+GB of RAM

so after it runs in 20 threads on 20 cores there's still about 10MB L3 cache and half the threads available. so I can run something like PASC miner there without actually increasing consumption:)

but then randomx performance gets degraded:

philipma1957

legendary

Activity: 4326

Merit: 8950

'The right to privacy matters'

Quote from: mstrozier on October 24, 2019, 09:06:48 PM

Thanks, I'll add the issue to the github. I've tried the two suggestions given, increasing the swap size and gcc 7+ but still getting the same error.

Has anyone gotten a rasp pi 4 to work?

mstrozier

member

Activity: 114

Merit: 151

Thanks, I'll add the issue to the github. I've tried the two suggestions given, increasing the swap size and gcc 7+ but still getting the same error.

herrdrone

sr. member

Activity: 633

Merit: 267

Quote from: mstrozier on October 24, 2019, 10:10:09 AM

Quote from: herrdrone on October 24, 2019, 07:52:23 AM

Quote from: philipma1957 on October 24, 2019, 07:45:54 AM

WHEN DOES THIS MINING START!!!!

On november 30, details: https://np.reddit.com/r/Monero/comments/d884zt/preliminary_information_thread_regarding_the/

In the meantime, you can try testnet pool like https://rx.minexmr.com/ with XMRig https://github.com/xmrig/xmrig/releases

Morning. I tested the RandomX bench on both a Raspberry Pi 3 b+ and Pi 4 and submitted results several weeks ago. I wanted to test out both units on the testnet pool however running into issues installing XMRig. From the github I downloaded the source, built the build folder, changed directory into the build folder then ran cmake .. and once that completed ran make

make ran into the following errors. Any ideas? I may test further on a fx8350 linux mint box that I have up running or my Ryzen 2600. But was curious on the pi's usage

Errors are below:

Code:

In file included from /home/pi/mining/test/src/crypto/cn/SSE2NEON.h:123,
                 from /home/pi/mining/test/src/crypto/cn/soft_aes.h:31,
                 from /home/pi/mining/test/src/crypto/cn/CryptoNight_arm.h:34,
                 from /home/pi/mining/test/src/crypto/cn/CnHash.cpp:35:
/usr/lib/gcc/arm-linux-gnueabihf/8/include/arm_neon.h: In function ‘__m128i _mm_set_epi32(int, int, int, int)’:
/usr/lib/gcc/arm-linux-gnueabihf/8/include/arm_neon.h:10369:1: error: inlining failed in call to always_inline ‘int32x4_t vld1q_s32(const int32_t*)’: target specific option mismatch
 vld1q_s32 (const int32_t * __a)
 ^~~~~~~~~

make[2]: *** [CMakeFiles/xmrig.dir/build.make:1532: CMakeFiles/xmrig.dir/src/crypto/cn/CnHash.cpp.o] Error 1
make[1]: *** [CMakeFiles/Makefile2:73: CMakeFiles/xmrig.dir/all] Error 2
make: *** [Makefile:84: all] Error 2

I ran this on 2 different Raspberry Pi 4's. Both on 64 bit O/S Raspbian. One of them with the kernal update to 64 as shown below

Code:

Linux Goober-Pi4 4.19.75-v8+ #1270 SMP PREEMPT Tue Sep 24 18:59:17 BST 2019 aarch64 GNU/Linux

2nd Pi 4 where the kernal hasn't been updated to 64 (this one is a controller for R606 / Newpac miners):

Code:

Linux Pi-Miner 4.19.66-v7l+ #1253 SMP Thu Aug 15 12:02:08 BST 2019 armv7l GNU/Linux

Thanks

Hey, doing some reading I suggest you to increase swap area (https://www.linuxtechi.com/extend-swap-space-using-swap-file-in-linux/) and use gcc 7+. And try again.

Yet, It is better if you add this issue on XMrig's github to get help: https://github.com/xmrig/xmrig/issues

Marvell2

full member

Activity: 1148

Merit: 132

Quote from: philipma1957 on October 24, 2019, 07:45:54 AM

WHEN DOES THIS MINING START!!!!

just mine loki on miner rocks i think , its using a randomx fork, ive been on it for the last month using it to tune my 4 rigs

WhyMe

sr. member

Activity: 661

Merit: 250

Quote from: Ursul0 on October 24, 2019, 07:36:16 AM

2*Xeon E5 2660 v2 give 7330h/s on RandomX with 1 thread per core (L3 cache = 25MB*2slots)
setting higher thread count will degrade performance of paired(on a single core) threads

2* Xeon E5645 ( total is 12c/24t @2.4GHz )
Default setting is 12t and give 2220h/s. Setting other than 12t gives lower hashrate.
6t = 1680h/s
8t = 2030h/s
16t = 1920h/s
20t = 1640h/s ( and with 20t server is already 98-99% cpu usage )

mstrozier

member

Activity: 114

Merit: 151

Quote from: herrdrone on October 24, 2019, 07:52:23 AM

Quote from: philipma1957 on October 24, 2019, 07:45:54 AM

WHEN DOES THIS MINING START!!!!

On november 30, details: https://np.reddit.com/r/Monero/comments/d884zt/preliminary_information_thread_regarding_the/

In the meantime, you can try testnet pool like https://rx.minexmr.com/ with XMRig https://github.com/xmrig/xmrig/releases

Morning. I tested the RandomX bench on both a Raspberry Pi 3 b+ and Pi 4 and submitted results several weeks ago. I wanted to test out both units on the testnet pool however running into issues installing XMRig. From the github I downloaded the source, built the build folder, changed directory into the build folder then ran cmake .. and once that completed ran make

make ran into the following errors. Any ideas? I may test further on a fx8350 linux mint box that I have up running or my Ryzen 2600. But was curious on the pi's usage

Errors are below:

Code:

In file included from /home/pi/mining/test/src/crypto/cn/SSE2NEON.h:123,
                 from /home/pi/mining/test/src/crypto/cn/soft_aes.h:31,
                 from /home/pi/mining/test/src/crypto/cn/CryptoNight_arm.h:34,
                 from /home/pi/mining/test/src/crypto/cn/CnHash.cpp:35:
/usr/lib/gcc/arm-linux-gnueabihf/8/include/arm_neon.h: In function ‘__m128i _mm_set_epi32(int, int, int, int)’:
/usr/lib/gcc/arm-linux-gnueabihf/8/include/arm_neon.h:10369:1: error: inlining failed in call to always_inline ‘int32x4_t vld1q_s32(const int32_t*)’: target specific option mismatch
 vld1q_s32 (const int32_t * __a)
 ^~~~~~~~~

make[2]: *** [CMakeFiles/xmrig.dir/build.make:1532: CMakeFiles/xmrig.dir/src/crypto/cn/CnHash.cpp.o] Error 1
make[1]: *** [CMakeFiles/Makefile2:73: CMakeFiles/xmrig.dir/all] Error 2
make: *** [Makefile:84: all] Error 2

I ran this on 2 different Raspberry Pi 4's. Both on 64 bit O/S Raspbian. One of them with the kernal update to 64 as shown below

Code:

Linux Goober-Pi4 4.19.75-v8+ #1270 SMP PREEMPT Tue Sep 24 18:59:17 BST 2019 aarch64 GNU/Linux

2nd Pi 4 where the kernal hasn't been updated to 64 (this one is a controller for R606 / Newpac miners):

Code:

Linux Pi-Miner 4.19.66-v7l+ #1253 SMP Thu Aug 15 12:02:08 BST 2019 armv7l GNU/Linux

Thanks

herrdrone

sr. member

Activity: 633

Merit: 267

Quote from: philipma1957 on October 24, 2019, 07:45:54 AM

WHEN DOES THIS MINING START!!!!

On november 30, details: https://np.reddit.com/r/Monero/comments/d884zt/preliminary_information_thread_regarding_the/

In the meantime, you can try testnet pool like https://rx.minexmr.com/ with XMRig https://github.com/xmrig/xmrig/releases

philipma1957

legendary

Activity: 4326

Merit: 8950

'The right to privacy matters'

WHEN DOES THIS MINING START!!!!

Metroid

sr. member

Activity: 2142

Merit: 353

Xtreme Monster

in the end profit will determine its value, good seeing ryzen 3xxx doing good, this is good, something that a cpu is more useful than a gpu.

3900x or 3950x owners will be happy with this.

Topic: RandomX Benchmarks! https://monerobenchmarks.info - page 8. (Read 6455 times)