Topic: PhoenixMiner 6.2c: fastest Ethereum/Ethash miner with lowest devfee (Win/Linux) - page 310. (Read 785074 times)

2Gb,4Gb DAG fix, need atleast 1 GPU large enough to generate DAG.

Recently realized in my farm that my 470 4Gb rigs, which were all 6 card setups at the time, were not generating the dag any longer for ETC and then soon after ETH dag approached whatever level it was that pushed my system over the limit. After a day of wondering what to do, I realized a 470 rig I have with 1 8Gb 580 in it was working fine with the other 5 cards being 470 4Gbs. I took apart one of my 580 8Gb rigs and replaced a card on one of the 470 rigs. Mined perfectly fine. Did this to the whole farm and split the gpus essentially so every 4Gb rig has atleast 1 8Gb card in it. As for now this works no issues. I have a p104-100 rig with 1 470 4Gb in it and same thing, generated dag and mines.

I tested this on 2Gb R7 200 gpus and works perfectly fine. Card generates the dag as long as the system has one available card that is able to fit the dag on it. This is identical to how the 470 rigs are operating.

The same goes for 470 4Gbs with rx 5700s and the same with the R7 200s with the rx 5700s.

It sounds like 4GB isn't enough for the DAG since the current DAG is 3.27GB for ETH or 3.34GB for ETC.
RedPandaMining recently posted a video on his Youtube Channel.
https://www.youtube.com/watch?v=xRIX5mpVntk

your discussion is seriously weird. it is impossible to load >3gb array into <3 GPU memory whatever tricks you may perform.

DAG  - directed acyclic graph

"The Ethash algorithm expects the DAG as a two-dimensional array of uint32s (4-byte unsigned ints), with dimension (n × 16) where n is a large number. (n starts at 16777186 and grows from there.) Following the magic number, the rows of the DAG should be written sequentially into the file, with no delimiter between rows and each unint32 encoded in little-endian format."

"...calculating the PoW (Proof of Work) requires subsets of a fixed resource dependent on the nonce and block header. This resource (a few gigabyte size data) is called a DAG. The DAG is totally different every 30000 blocks (a 100 hour window, called an epoch) and takes a while to generate."

no. the situation seems very much similar to RX5(4)70 & RX5(4)80 - 10% more shaders making no difference. at 1200MHz core and 900MHz mem, @750MV it produces 50-51MH and consumes about 100W

I got this error
2019.10.24:07:22:23.857: main CUDA version: 10.0, CUDA runtime: 8.0
2019.10.24:07:22:23.986: main No OpenCL platforms found
2019.10.24:07:22:23.986: main Available GPUs for mining:
2019.10.24:07:22:23.986: main GPU1: P106-100 (pcie 1), CUDA cap. 6.1, 5.9 GB VRAM, 10 CUs
2019.10.24:07:22:23.986: main GPU2: P106-100 (pcie 2), CUDA cap. 6.1, 5.9 GB VRAM, 10 CUs
2019.10.24:07:22:23.986: main GPU3: P106-100 (pcie 3), CUDA cap. 6.1, 5.9 GB VRAM, 10 CUs
2019.10.24:07:22:23.986: main GPU4: P106-100 (pcie 5), CUDA cap. 6.1, 5.9 GB VRAM, 10 CUs
2019.10.24:07:22:23.986: main GPU5: P106-100 (pcie 6), CUDA cap. 6.1, 5.9 GB VRAM, 10 CUs
2019.10.24:07:22:23.986: main GPU6: P106-100 (pcie 7), CUDA cap. 6.1, 5.9 GB VRAM, 10 CUs
2019.10.24:07:22:23.986: main GPU7: P106-100 (pcie , CUDA cap. 6.1, 5.9 GB VRAM, 10 CUs
2019.10.24:07:22:23.998: main NVML library initialized
2019.10.24:07:22:24.104: main Nvidia driver version: 431.60
2019.10.24:07:22:24.106: main NVAPI error in NvapiWrapper.c:101 : -6
2019.10.24:07:22:24.106: main Unable to load NvAPI: error 3
2019.10.24:07:22:24.606: main Eth: the pool list contains 2 pools (2 from command-line)
2019.10.24:07:22:24.606: main Eth: primary pool: ssl://eu1.ethermine.org:5555
2019.10.24:07:22:24.606: main Starting GPU mining
2019.10.24:07:22:24.606: main GPU1: unable to init NvAPI with bus ID 65536
2019.10.24:07:22:24.607: main GPU2: unable to init NvAPI with bus ID 131072
2019.10.24:07:22:24.607: main GPU3: unable to init NvAPI with bus ID 196608
2019.10.24:07:22:24.607: main GPU4: unable to init NvAPI with bus ID 327680
2019.10.24:07:22:24.607: main GPU5: unable to init NvAPI with bus ID 393216
2019.10.24:07:22:24.607: main GPU6: unable to init NvAPI with bus ID 458752
2019.10.24:07:22:24.607: main GPU7: unable to init NvAPI with bus ID 524288
2019.10.24:07:22:24.752: wdog Starting watchdog thread

Tried -mi8, no luck.

Claymore runs for a week without issues on the same clocks too.

Do you guys still have problems with RX5700 and Nicehash?

Do you see any performance difference between RX5700 and RT5700XT?

Reduce whatever you like from mining intensity -mi (I recommend this) , -mclock or power limit -powlim so the performance to drop with 0.5% (0.25MH/s). This will reduce the overheating of memory and possibly will stop the crashes.

Shap Pulse XT and Shap Nitro XT

Shap Pulse XT and Shap Nitro XT

Yes, I did. Same performance, no difference for me.

Anyone tried new AMD 19.10.2 driver with PM?

It would be awesome if it worked on 3gb Nvidia cards, but from what I have seen its only a fix for AMD cards with 4gb and below running on windows. I setup a windows box with a GTX 1070 and a couple gtx 3gb 1060s with no luck, then I tried a mixed mode setup with an 8gb AMD card as the main GPU with no luck as well. I figured it would not work with Nvidia cards below 3 gb, but I wanted to try it out for kicks. If anyone else tried and had luck let me know.

This is the most uselss comm i´ve read in the last 4 years ))))) Especially in ´´ No one is getting reacher with 3-5 MH in plus at one card. ´´´. you mean , doesnt matter on my 324 GPU 580 / 590 and a few Vega 56 ?! Get out and look for a job or delivery pizza in your town hommie i used claymore and phoenix . with CM i have 3 mhs more at the same wattage , with phoenix less dev fee

Who is the vendor of your RX 5700? What is the exact model? Is your 5700 XT or not? What overclock settings are you using?

Guys why 5700 gives the same hashrate 52mgh at 1300 core and at 2000 core frequency?