Welcome to ExtremeHW
Welcome to ExtremeHW, register to take part in our community, don't worry this is a simple FREE process that requires minimal information for you to signup.
Registered users can:
- Start new topics and reply to others.
- Show off your PC using our Rig Creator feature.
- Subscribe to topics and forums to get updates.
- Get your own profile page to customize.
- Send personal messages to other members.
- Take advantage of site exclusive features.
- Upgrade to Premium to unlock additional sites features.
tictoc
-
Posts
647 -
Joined
-
Last visited
-
Days Won
43 -
Feedback
0%
Content Type
Forums
Store
Events
Gallery
Profiles
Videos
Marketplace
Tutorials
Everything posted by tictoc
-
Actually, nothing at the moment. I didn't realize that the TC is going to kick back off August 1st, so for now we are all good.
-
I don't remeber the last time I logged in to Discord, so thanks for the heads up. The Foldaholics will be all sorted out to start our reign of destruction on August 1st.
-
10-4. Hit me up if you're interested in some no name 16GB 3200 ecc udimms. @firedfly are you still interested in temping with your Quadro or any other GPU?
-
@axipher @Avacado Now that the forum is revived, I just wanted to see where we were at. I have a 980 that I can slot into Wildcard if the 5700 is still going to be up and folding. Let me know what's up and then I'll let @BWG know what we are all folding on.
-
My points are going to go to zero for the next week. I'm moving the Radeon VII off F@H and testing a few different GPUs. I don't want my points to be all over the place, so that passkey will just be offline for a bit.
-
Unlike mining, there is quite a bit of communication between the CPU/Memory and the GPU while running a WU. It is not an extremely bandwidth intensive GPU compute app, but once you go below PCIe 3.0 x8, with faster GPUs, the performance will drop.
-
I don't have an x570|Ryzen 5000 setup, but first thing I would try is swapping your NVMe drive into one of the chipset slots. If that doesn't work, then I would look at running a SATA SSD, which will be more than adequate for a folding machine. I went round and round getting 3 GPUs running on my x370|1700 machine. I was able to get it working, but that was on three 7970s that were not running anything that was constrained by the PCIe bandwidth.
-
I have a few things that could slow that down.
-
This made me look at my EHW stats. Looks like I just passed 500M for the team. I also saw that @BWG and @firedfly were threats. In about 11 months. ?
-
-
IN (out on the prize draw) Q_Cruncher -
@Avacado No worries if you need to take a break. Do you think you'll be back for the "official" start in July or is there no ETA on getting your rig back up and running? @firedfly Are you still planning on dedicating the 3070 to 24/7 folding? If Avacado is going to be on the bench for a bit, then we'd be happy to have you on the team with whatever GPU you want to fold on.
-
I don't think so, it looks like 1.7.3 released 5/14/20 is the latest.
-
Not sure if anyone is still using this or following this thread, but the "New Version?" button is a broken link. @ENTERPRISE maybe we could redirect the broken link to here. Broken link address: https://extremehw.net/forum/amd/general/6090-dram-calculator-for-ryzen™-news-and-discussions?p=6232#post6232
-
Right on. I'm still not positive what I'll be folding on. My schedule is pretty busy at the moment, but I'll have some time in the next few weeks to finalize what I'll be folding on.
-
I took a day off from folding on the Radeon VII, while I sorted a few things out on my workstation. It's back up and running now. I rolled back to the old closed-source amdgpu-pro driver to do some testing on some non-F@H work, and much to my surprise, the current WU that I'm running (beta WU 18201) is seeing a massive performance improvement running on the old driver. amd-staging-drm-next kernel + ROCm 4.2 = 2M ppd 5.12.9 kernel + OpenCl from amdgpu-pro 20.30_1109583-1 = 3.1M ppd
-
You could put a USB 3.0 card in the PCIe 2.0 x4 slot, and use that to drive your front monitors. https://vantecusa.com/products_detail.php?p_id=125&p_name=
-
I would be fine with using the All OS average, and then doing a week long test, just to make sure that everything is in line. The only real issue is that if there aren't any recent results then the average on LARS is probably going to include a bunch of pre-CUDA tasks. On average the old OpenCL NVIDIA tasks were at least 25-30% less PPD, compared to the current CUDA tasks. I'm not 100% sure on the above, since I have no idea how long the LARS database has been up and running.
-
I have not had time to do much testing. Although it looks like I should run some low spec hardware for max points. I have a high clocking AMD 5770, that would probably give me twice the adjusted points compared to any of my other GPUs. In the little testing that I've done it looks like my 980 classy, that won a few times in the old TC, would also put out fairly decent points. In order to really figure it out I need to free up my test bench. Hopefully that will happen in the next couple of days. I have the chassis for my server stripped down, and I should be able to get that board off the test bench this week.
-
The ASRock Rack board has a BMC with onboard VGA, so no issues there unless the primary video is set from auto or onboard to PCIe in the BIOS. My Taichi board, on the release BIOS, would hang on post with no GPU installed. I'm not sure if I ever tried it on a more recent BIOS. There is really no technical reason for failing to post without a GPU, except for shoddy UEFI programming.
-
No issues on the ASRock boards that I have. The three boards will happily run headless, as long as there is a GPU in the primary slot.
-
No 960. The only Maxwell I have is a 980.
-
Not sure if it's just my browsers, but the "Unmultiplied points" tooltip is showing up far from the cursor. Runnning latest Firefox and Chromium in Linux. Firefox: Chromium:
-
Back up and running with a Radeon VII. I'll be testing out some other GPUs this week, so I might be changing my TC card for June.
-
@Avacado You could mine other coins on the M40, but not ETH at the current DAG (4GB). Maxwell only has 1GB of addressable TLB, and that is why you see low hash rates.
