teamanandtech.org

Target corona_PLpro_v1 is now at ≈91 %. Maybe the current ≈9,600 tasks in progress make up the missing 9 %. ( https://www.sidock.si/sidock/server_status.php ) We will soon finish the second target SARS-CoV-2 PL -Pro as well. I'm not sure what this means, but for now I interpret this as "We will soon...

If they make its TBP ≈20 W, not 30 W like reported, then it could be a 1:1 GT 710 replacement, which they then could phase out. (Did the Kepler based GT 710 already replace a prior Fermi based eponymous GT 710? — Edit, nope, the Fermi based one was called GT 705.)

Interesting read on Covid-19 variants. link Coincidentally, just now I heard a radio interview, recorded January 14, with an epidemiologist on the topic of the current mutations. It was said that there is a suspicion that immune escape is beginning now with the Brazilian variant, but it is not clea...

<r>Correct. You can enter as many <URL url="https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/admin-guide/kernel-parameters.txt"><s>[url=https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/admin-guide/kernel-parameters.txt]</s>ker...

While I never profiled its effects, I have mitigations off on all of my DC-dedicated computers.
It was a small nuisance to figure out how to set the GRUB config accordingly in Mint.
In OpenSuse, there is a dialogue for this in their central system management GUI.

January 1 --> 8 .............. 0 --> 27 % completion (average: 3.9 %/day)
January 8 --> 11 .......... 27 --> 44 % completion (average: 5.7 %/day — 10 days to go at this rate with this target)

From what I remember, the now long submerged Acoustics@home had better performance under Windows than under Linux. They linked the Windows build against a heavily optimized non-free math library, while the Linux build had to make do with a version which lacked all the optimizations.

HD 6450 ....... TeraScale 2 (Northern Islands) Caicos Pro ....... 625 MHz ..... 40 nm TSMC BTW, EPYC: As you probably know, I preferred to build dual-socket systems. Benefit: Half the amount of mainboards, PSUs, disks, network interfaces, BMCs/GPUs (and potentially: cases) compared to single socket ...

Some hardware specs: GT 1030 ............ Pascal GP108-300-A1 ................... 1228 MHz ..... 14 nm Samsung GTX 750 Ti ....... Maxwell GM107-400-A2 ................ 1020 MHz ..... 28 nm TSMC GT 730 .............. Kepler 2.0 GK208B ........................... 902 MHz ..... 28 nm TSMC GT 710 .........

Baseboard Management Controllers (BMCs) are not only low-power video cards, they also allow for remote management, even down to remote access to the BIOS. Downside: They don't come as add-in boards; you need to buy a mainboard with them. Related news: https://www.anandtech.com/show/16349/the-asrock-...

@Skillz you should get an SSL cert. Yeah I need to do some research on how to get that done. Haven't had time to do anything for a while now. If it requires $, let me know and I'll pay for it. Yes you can pay for one at like your domain provider but you can also get free ones so I would do that if ...

The last test completed now. I obtained the average of all task run times, calculated PPD and PPD/W from them, and edited them into the previous post. Conclusion: Performance of the GPU application is usable, but not exciting. One GTX 1080 Ti is equivalent to 19 Xeon E5 v4 threads, or to 14 EPYC Rom...

One more thing which I forgot to ask: Did you run a 32-bit Linux or a 64-bit Linux in the VM? BTW, I browsed top_hosts.php just now and looked at the reported GFLOPS of "Get Decic Fields 4.00" of some computers with same CPU but Windows versus Linux. The GFLOPS were all over the place. Therefore, to...

Personal NumberFields Sprint, Eleven hours later: The Beast (Windows): 823,500 - 778,625 = 44,875 epyc-bunker-1 (Mint 20): 177,375 -151,375 = 26,000 (Oracle Virtualbox) 1 : 0.58 How well are Windows + VirtualBox balancing the load on a 64-thread SMT machine? If you have the time, try a similar test...

I just released the new app version for Linux Nvidia Cuda. I tested it on 100 WUs and it averaged about 33% faster over the previous Cuda version. I am now making these same improvements to the openCL versions. If all goes well, I should have new apps for Windows and AMD cards within a week. ( sour...

BTW, this is old news (mid December). Gerasim's RSS feed apparently doesn't contain publishing time stamps, and therefore the posting date is from when this news was fetched by the http://teamanandtech.org site code. I'll need to keep an eye on this feed. If it just clutters our forum, I shall disab...

For performance comparisons, the fact that NumberFields is using fixed credits currently is nice. But the run time variations are huge. On a high-core-count CPU, half a day of testing time is certainly appropriate, but more than that may be necessary on low-core-count CPUs. For quicker testing with ...

BTW, idle power consumption of Turing and Ampere is higher again than Pascal — as far as I have read. Also, I concede that 1070/1080/1080Ti are a little off-topic here, as you specifically looked for GPUs which don't give you the idea to run BOINC on them. My methodology for DC used to be having low...

Machine translation: Guys, there is something to look at here: https://classic.boincstats.com/en/stats/64/user/detail/3050/charts Excellent credits from fellow PDW. Is not it? 74 million a day ... p.s. I think there was some mistake here. How can we doubt the honesty of our dear comrade PDW? And by ...

1080Ti is generally at circa 8 W when idle (in Linux, measured by the board sensor). Or maybe more like 9 or 10 W for a GPU which displays an idle Cinnamon desktop, and 7 W for an unused secondary GPU.

If I remember correctly, 1070 and 1080 had the same idle consumption.

How well are Windows + VirtualBox balancing the load on a 64-thread SMT machine? If you have the time, try a similar test while SMT is disabled in the BIOS. It is safe to say that virtualization comes with a mininum of performance hit when set up accordingly. The business of cloud providers depends ...

I went ahead and created two topics sub-forums. But if there are better ideas for the titles and descriptions, I'll be glad to apply any. Edit: when I moved several threads into the new forums, the posts were marked as unread again. For me at least. If they show up as unread to you too, then I am ap...

TLDR: remote code execution vulnerability fixed in FAHClient v7.6.20 vulnerability existed if you ran client remote control over a public network generally, DON'T use FAHClient remote control over a public network — restrict it to your LAN, or use a secure tunnel a.k.a. virtual private network (I ju...

I created this "Forum Issues" topic now and moved some threads from "The Camp Fire" in here. I also took the liberty to move some other threads from The Camp Fire into "Distributed Computing News" and "The Market", and extended the description of "The Market" a little. Currently we have got a lot of...

I don't have a comparison between Haswell-EP and Broadwell-EP, which seems to be the point here. But there is some other data if you copy+paste Some arbitrary performance measurements into the search box at the top and hit the magnifying glass: The 14-core E5-2690 v4 gives 0.7 of the performance of ...

crashtech , good to hear that at least the CPUs seem to be in working condition. But it's strange and a bummer that the upgrade of the Asus board didn't work. I've got no experience with Asus server boards myself. I have a monoculture of Supermicro boards here. I did have painful experiences with A...

I don't know. Intel certainly test it when they manufacture these.

Maybe the BIOS version does not bring up the QPI links correctly.
Or maybe one of the CPUs has a damaged QPI link.

I suppose you have checked the reported model number when you booted with only one socket populated, to verify that these are genuine. Are they production samples, or ES?

Last year's challenge consisted of the entire [H]ard|OCP team (no sign-up), versus TeAm AnandTech members who signed up. Maybe have [H] members sign up too, and then make 3 teams. Edit: That'd be a lot more work for the stats master though.