PhoenixMiner 6.2c: fastest Ethereum/Ethash miner with lowest devfee (Win/Linux)

[Semtex72]

Performance Video in my testing so far: 481mh @ 1062w for a 8x 5700 XT rig. https://twitter.com/BitsBeTrippin/status/1174169921250631682

But probably going to settle for 474mh @ 1058, as few of the cards not showing 100% stability with settings. Once I see it clean for 24h will post how I got there in video.

Again, please share your stable settings checked vs the pool

I run like this 100% stable:

-tt 58 -cvddc 750 -cclock 1300 -fanmin 0 -mclock 910 -amd -mode 1 -clKernel 1 -clNew 1

[Semtex72]

I own now 2x Sapphire 5700XT.

I have used Adrenaline 1.9.2.2 and classic setting posted on this topic; -tt 62  -cclock 1300 -cvddc 800 -mclock 900 -hstats 2 it gives 52,x MH but it fails in minutes.

I have tried lower or bigger -cvvdc between 750 and 900mV, even lower/higher -tt, it fails randomly showing 0.00MH.

I tried also to keep memory stock (875MHz), it fails randomly, same problem and miner stays locked, mouse too, just hard reset works.

When I have removed -hstats 2 command, keeping just  -tt 62  -cclock 1300 -cvddc 800 -mclock 900, it worked fine 12 hours but I just tried last sollution: I have deleted all commands -tt 62  -cclock 1300 -cvddc 800 -mclock 900 -hstats 2 from ,bat file and I have used just Msi Afterburner 4.6.2 beta 2           https://www.guru3d.com/files-details/msi-afterburner-beta-download.html        wich full controls power limit, voltage and frequency, now I left voltage unattended (1187mV), GPU frequency stock, memory increased at 900MHz and power limit down to -43 or even -44%.

Works flawless.

It looks as the much commands gived to the miner (tt 62  -cclock 1300 -cvddc 800 -mclock 900 -hstats 2), especially -hstats 2,  gives driver error (atimtag.dll) and locks the miner...

So, I keept settings as in picture attached and .bat file as simple, without extra commands ex:  PhoenixMiner.exe -pool ssl://eu1.ethermine.org:5555 -pool2 ssl://us1.ethermine.org:5555 -wal your wallet adress.rig name, and is rock solid.

https://i.postimg.cc/RhtKXjtg/Best-settings.jpg

Seems to be these issues are related to nicehash (even if that is not mentioned here).

[riqtroy]

Need help.
Everything was fine until one week ago, when suddenly I'v start getting these errors:

12 GPU: "CUDA error in CudaProgram.cu:407 : out of memory" with "-eres" more than "0" (default 2);
12 GPU: "CUDA error in CudaProgram.cu:195 : out of memory" on 92% DAG with "-eres 0";
11 GPU: same as 12 GPU
10 GPU: "CUDA error in CudaProgram.cu:407 : out of memory" with "-eres" more than "0" (default 2);
10 GPU: Works fine with "-eres 0";

also I don't remember when, but sometimes I get:
"CUDA error in CudaProgram.cu:173 : out of memory",
"CUDA error in CudaProgram.cu:204 : out of memory".

RIQ:
Windows 10 64 (can't install win7 bcuz then motherboard won't support all gpu's)
mb: Asus B250
cpu: G4560
gpu: 1050TI x 12
ram: 8GB
ssd: 128GB (virtual memory set to 98304mb)

[riqtroy]

I have problem with error in CudaProgram, can not continue mining.
Its 6x GTX 1050ti rig, Windows 10.
Any help?

2019.09.23:13:13:10.914: GPU5 CUDA error in CudaProgram.cu:204 : out of memory (2)
2019.09.23:13:13:10.914: GPU5 GPU5: Failed to initialize search buffers: out of memory
2019.09.23:13:13:10.914: GPU5 GPU5 initMiner error: Unable to initialize CUDA miner
2019.09.23:13:13:10.917: GPU6 CUDA error in CudaProgram.cu:204 : out of memory (2)
2019.09.23:13:13:10.917: GPU6 GPU6: Failed to initialize search buffers: out of memory
2019.09.23:13:13:10.917: GPU6 GPU6 initMiner error: Unable to initialize CUDA miner
2019.09.23:13:13:10.920: GPU4 CUDA error in CudaProgram.cu:204 : out of memory (2)
2019.09.23:13:13:10.920: GPU4 GPU4: Failed to initialize search buffers: out of memory
2019.09.23:13:13:10.920: GPU4 GPU4 initMiner error: Unable to initialize CUDA miner
2019.09.23:13:13:10.933: GPU3 CUDA error in CudaProgram.cu:204 : out of memory (2)
2019.09.23:13:13:10.933: GPU3 GPU3: Failed to initialize search buffers: out of memory
2019.09.23:13:13:10.933: GPU3 GPU3 initMiner error: Unable to initialize CUDA miner
2019.09.23:13:13:11.232: GPU1 GPU1: DAG  88%

Have you tried to set "-eres 0"? It helps me with 10 gpu's or less

[PhoenixMiner]

Because of the GTX1050Ti (and other 4 GB Nvidia cards) problems under Windows, here is a special release that will solve the issue for a few more weeks. That's the best that can be done, because Nvidia drivers just won't allow to allocate enough VRAM even when it is available. Note that if you don't have 4 GB Nvidia cards, there is no point to use the 4.7a - there are only some small fixes for rare bugs. You can download PhoenixMiner 4.7a from here:

(MEGA links are no longer active)

If you want to check the integrity of the downloaded file, please use the following hashes (you need the last file PhoenixMiner_NVRTC_Windows.zip only if you want to mine BCI with Nvdia cards under Windows):

Code:

   File: PhoenixMiner_4.7a_Windows.zip
    ===================================
   SHA-1: c4448eef92538ff9745a8cd6a79e9f1dbd9d3f20
 SHA-256: 18077c9e18b66a43add6f65eceaad5c56359dffec3e5a0b2bdf609fb8cbb1340
 SHA-512: 45b87fa6f6e80b616bbf97ef02d572fac499660d4b8e854bc3624df6682276aae4341dfe24c04d460d5a44f1c986f6df54ad909ff03861fb4c656d21e716d324


    File: PhoenixMiner_NVRTC_Windows.zip
    ====================================
   SHA-1: ff6fa5e018adbd52caf631c42b7c2fac7ce48a51
 SHA-256: 8087757169405d51ea8ba818347fb05d0450aef985c29272165070346eb5a54a
 SHA-512: 7b2d832f7f40578bb1f501d5174467f5ae06612e601dab769fd56d39da48a471b18c6373435a485155f70fec4017d8378797bf1e1dfe5d62fee30fa6a1d992c4

Here are the new features for this release:

• Reduced VRAM usage for Nvidia cards on Windows to allow slightly longer operation of 4 GB cards under Windows 10 (few weeks to a month, then the problem will appear again)
• Other small improvements and fixes.

Remember to add -rvram -1 -eres 0 to your command-line or config.txt file in order to lower the memory usage as much as possible.

[PhoenixMiner]

Here are some notes about the Windows 10 out of memory problems:

• While they are present both on AMD and Nvidia GPUs, the problem with Nvidia is much worse. The 4 GB AMD GPUs will work without problems under Windows 10 at least until DAG 350, while the best that can be expected from Nvidia is DAG 293. And this is not entirely Windows 10 fault because even under Linux the 1050Ti won't be able to go beyond DAG 340 or a little higher.
• The first thing to do after encountering out of memory error is to add -rvram -1 -eres 0 to your command-line or config.txt file in order to lower the memory usage as much as possible.
• Again, there is no point to upgrade to PhoenixMiner 4.7a if you don't have 4 GB Nvidia cards and there are no out of memory problems

[riqtroy]

Here are some notes about the Windows 10 out of memory problems:

• While they are present both on AMD and Nvidia GPUs, the problem with Nvidia is much worse. The 4 GB AMD GPUs will work without problems under Windows 10 at least until DAG 350, while the best that can be expected from Nvidia is DAG 293. And this is not entirely Windows 10 fault because even under Linux the 1050Ti won't be able to go beyond DAG 340 or a little higher.
• The first thing to do after encountering out of memory error is to add -rvram -1 -eres 0 to your command-line or config.txt file in order to lower the memory usage as much as possible.
• Again, there is no point to upgrade to PhoenixMiner 4.7a if you don't have 4 GB Nvidia cards and there are no out of memory problems

Hello, PhoenixMiner!
So basically 1050ti can live under Windows 10 only until 293 which is 22 october 2019 or 8790000 block?
And the best way to go until the end with 1050ti is Linux? No way to make 383 even with Linux?
ps Thank you for 4.7a, 12 gpu's works fine.

[jpmonster]

Here are some notes about the Windows 10 out of memory problems:

• While they are present both on AMD and Nvidia GPUs, the problem with Nvidia is much worse. The 4 GB AMD GPUs will work without problems under Windows 10 at least until DAG 350, while the best that can be expected from Nvidia is DAG 293. And this is not entirely Windows 10 fault because even under Linux the 1050Ti won't be able to go beyond DAG 340 or a little higher.
• The first thing to do after encountering out of memory error is to add -rvram -1 -eres 0 to your command-line or config.txt file in order to lower the memory usage as much as possible.
• Again, there is no point to upgrade to PhoenixMiner 4.7a if you don't have 4 GB Nvidia cards and there are no out of memory problems

This is not working for me:(

[damNmad]

Hello there,

I'm getting 'Bad -coin value: qkc' error while mining Quark (qkc) coin?

I'm using '-coin qkc'

[kjang0517]

The fan control does not apply (Ubuntu 18.04.3, Titan V)

[Nightworker88]

@ damNmad

just use auto ,  eg.: ........-cdmport 3332 -coin auto -retrydelay 5 -timeout 2880 -amd.........


working perfect

[aarts]

Is there option to do miner command "Z"(autotune) in intervals? Like every 3 hours .
 If difficult changes renewing GT option makes mining better.
https://hiveos.farm?ref=14664

[UnclWish]

Is there option to do miner command "Z"(autotune) in intervals? Like every 3 hours .
 If difficult changes renewing GT option makes mining better.
https://hiveos.farm?ref=14664

With the same core and memory clocks, and clKernel gt best parameter stays the same. It's no need to find it regural.

[aarts]

Yes all stays same. But it starts mining better if renew gt settings. Maybe its some connection issue then.

[zertor]

Here are some notes about the Windows 10 out of memory problems:

• While they are present both on AMD and Nvidia GPUs, the problem with Nvidia is much worse. The 4 GB AMD GPUs will work without problems under Windows 10 at least until DAG 350, while the best that can be expected from Nvidia is DAG 293. And this is not entirely Windows 10 fault because even under Linux the 1050Ti won't be able to go beyond DAG 340 or a little higher.
• The first thing to do after encountering out of memory error is to add -rvram -1 -eres 0 to your command-line or config.txt file in order to lower the memory usage as much as possible.
• Again, there is no point to upgrade to PhoenixMiner 4.7a if you don't have 4 GB Nvidia cards and there are no out of memory problems

Thanks, switched back to PhoenixMiner  

And this is not entirely Windows 10 fault because even under Linux the 1050Ti won't be able to go beyond DAG 340 or a little higher.[/li][/list]

340 vs 350 is not that much different compared to AMD
will be still usable for some time.

You say Linux,
how about Windows 7 compared to Linux?
I heard in there[Win7], there is no such issues like Win10
I.e: Could be Win7 to be alternative to Linux for that matter?

This is not working for me:(

paste the log file what happens exactly
are you using 4.7a version?
did you applied -rvram -1 -eres 0 too?
what OS and VGA drivers you use?
what's your virtual memory file size setup?
do you use another video card as primary display or you use the miners card for this?

[Xazax310]

I've flash my P104-100 to "8GB" bios but still runs out of ram when launching the DAG. guess it's just a BIOs trick?

[Ursul0]

Just curious here -  are there any plans to introduce/integrate memory straps into the miner like Claymore's v15 does?

[UnclWish]

Just curious here -  are there any plans to introduce/integrate memory straps into the miner like Claymore's v15 does?

It's allready done by author:

  -mt <n> VRAM timings (AMD under Windows only): 0 - default VBIOS values; 1 - faster timings; 2 - fastest timings.
     The default is 0. This is useful for mining with AMD cards without modding the VBIOS.

It's not memory modded memory straps, it's just setting options wich exists in Radeon Settings - timings section. But without vbios modding it can significant increase speed.

-rxboost option is absent in this moment... I hope Phoenix will add it in future too.

[Ursul0]

Just curious here -  are there any plans to introduce/integrate memory straps into the miner like Claymore's v15 does?

It's allready done by author:

  -mt <n> VRAM timings (AMD under Windows only): 0 - default VBIOS values; 1 - faster timings; 2 - fastest timings.
     The default is 0. This is useful for mining with AMD cards without modding the VBIOS.

It's not memory modded memory straps, it's just setting options wich exists in Radeon Settings - timings section. But without vbios modding it can significant increase speed.

-rxboost option is absent in this moment... I hope Phoenix will add it in future too.

interesting, thanks for the info! Never seen any timings settings in the drivers. I still use the best performing 18.6.1 and also without gui. will check it out later anyway.
Also regarding the Claymores "rxboost" I'd guess its REF and RFC that you can (among all other values) update with Amd mem tweak.  
Im hoping for Nvidia timings though:)

[UnclWish]

Just curious here -  are there any plans to introduce/integrate memory straps into the miner like Claymore's v15 does?

It's allready done by author:

  -mt <n> VRAM timings (AMD under Windows only): 0 - default VBIOS values; 1 - faster timings; 2 - fastest timings.
     The default is 0. This is useful for mining with AMD cards without modding the VBIOS.

It's not memory modded memory straps, it's just setting options wich exists in Radeon Settings - timings section. But without vbios modding it can significant increase speed.

-rxboost option is absent in this moment... I hope Phoenix will add it in future too.

interesting, thanks for the info! Never seen any timings settings in the drivers. I still use the best performing 18.6.1 and also without gui. will check it out later anyway.
Also regarding the Claymores "rxboost" I'd guess its REF and RFC that you can (among all other values) update with Amd mem tweak.  
Im hoping for Nvidia timings though:)

Be carefull, I didn't know will -mt work with old drivers or not...