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.
-
Posts
637 -
Joined
-
Last visited
-
Days Won
33 -
Feedback
0%
Content Type
Forums
Store
Events
Gallery
Profiles
Videos
Marketplace
Tutorials
Posts posted by tictoc
-
-
It's an EJ257, and the turbo cars (with the MLS gaskets) do not have the same head gasket issues that the NA cars have EJ251/53.
Generally speaking head gasket failure is not that common on the turbo motors. Ironically, I did blow a head gasket at about 130k miles, but that has been the only issue I've had with the car. It is probably a bit of a unicorn, since it's been pushing well over stock power for 230k miles.
-
-
Looks like I'm not in this yet.
Latest addition:
Unicomp Mini M
Daily driver | Filco Majestouch 2 TKL Camo | MX Brown | Keycaps: Blank Black PBT and Double Shot PBT Dark Gray w/ Black Legends
Work board | Cooler Master Quick Fire TK | MX Brown w/ dampers
On the KVM | Max Blackbird TKL | MX Brown
Spare board | Corsair K65 RGB Mini Black | MX Speed
- 1
- 1
-
-
I've been out since the beginning of the month (BOINC Pentathlon), but I'll have some gear back up an folding later this week.
- 1
-
Everything is up and running, and it will be getting a two week long stress test with the CPU and GPUs running 24/7, all out, for the BOINC Pentathlon.
Setting up the whole software stack for the server will have to wait, so for now it's just running a minimal headless Arch setup for crunching.
- 1
- 1
-
I think one of the bigger problems is that a large percentage of "media" companies are owned or run by the companies that are actually delivering the media. This is especially a problem when there are large swathes of the US outside of major cities that only have one or maybe two options for internet/TV service.
Companies like Comcast, AT&T, Verizon, and Charter are able to leverage their infrastructure to prop up their own media companies. This requires an obscene amount of oversight, since it would be easy for the companies that deliver the media to push the scales to favor their own in house media companies.
- 1
-
double post
-
20 hours ago, Sir Beregond said:
Now I am curious.
If you're curious, you can grab a copy and give it a go. https://archive.org/search.php?query=creator%3A"PeoplePC"
- 1
-
Nice work last month @axipher
I will not have my GPUs running 24/7 between now and the 19th. The BOINC Pentathlon has begun, and even though my team has gotten much smaller, I will be giving it all I've got. There will still be some folding time when my GPUs are not active. I'll be back to full time shortly after the 19th.
-
26 minutes ago, Diffident said:
At least on Linux it's using the proper forward slash instead of the uncomfortable backslash.
PowerShell is crazy in that it will display the path "properly" on Linux, but you can do whatever you want from the command line. The aggressive path normalization can break all sorts of things in a mixed Unix/Windows environment.
From the command line these are all the same and the shell doesn't care:
PS /> gci -Path home/tictoc/temp/testing/powershell -i *`(1`)* -r | ri
PS /> gci -Path home\tictoc\temp\testing\powershell -i *`(1`)* -r | ri
PS /> gci -Path home\tictoc/temp\testing/powershell -i *`(1`)* -r | ri
- 1
-
I imagine most of that performance lead evaporates against a 5995WX. IPC improvement of Zen 2 to Zen 3 is somewhere around 19% plus the higher stock clocks on a 5995WX.
The real question is how Sapphire Rapids Xeons will compare to current Milan-X CPUs and the Zen 4 Genoa chips. Depending on the workload, a 7773X is a pretty nice improvement over a 7763 (especially for a mid-process refresh), to say nothing of the upcoming Genoa chips and it's rumored monster L2 cache and 20% performance improvement.
- 1
- 2
-
Late to the show, but this is the kind of task that can be quickly accomplished from the CLI. More than likely this is what those GUI apps are using under the hood.
Since you're on windows we can use PowerShell with a few cmdlets. *Note* I am running this from PowerShell on Linux, but everything should be the same on Windows.
Here is a test scenario I created to accomplish this from the CLI.
List directories, sub-directories, and files in the current path:
SpoilerPS /home/tictoc/temp/testing/powershell> Get-ChildItem -Recurse Directory: /home/tictoc/temp/testing/powershell UnixMode User Group LastWriteTime Size Name -------- ---- ----- ------------- ---- ---- drwxr-xr-x tictoc tictoc 4/28/2022 19:00 4096 music drwxr-xr-x tictoc tictoc 4/28/2022 19:02 4096 music2 Directory: /home/tictoc/temp/testing/powershell/music UnixMode User Group LastWriteTime Size Name -------- ---- ----- ------------- ---- ---- -rw-r--r-- tictoc tictoc 4/28/2022 17:37 0 test (1).mp3 -rw-r--r-- tictoc tictoc 4/28/2022 17:33 0 test.mp3 -rw-r--r-- tictoc tictoc 4/28/2022 17:37 0 test1 (1).mp3 -rw-r--r-- tictoc tictoc 4/28/2022 17:36 0 test1.mp3 -rw-r--r-- tictoc tictoc 4/28/2022 17:39 0 test2 (1).mp3 -rw-r--r-- tictoc tictoc 4/28/2022 17:38 0 test2.mp3 -rw-r--r-- tictoc tictoc 4/28/2022 17:39 0 test3 (1).mp3 -rw-r--r-- tictoc tictoc 4/28/2022 17:38 0 test3.mp3 -rw-r--r-- tictoc tictoc 4/28/2022 17:39 0 test4 (1).mp3 -rw-r--r-- tictoc tictoc 4/28/2022 17:38 0 test4.mp3 -rw-r--r-- tictoc tictoc 4/28/2022 17:39 0 test5 (1).mp3 -rw-r--r-- tictoc tictoc 4/28/2022 17:38 0 test5.mp3 Directory: /home/tictoc/temp/testing/powershell/music2 UnixMode User Group LastWriteTime Size Name -------- ---- ----- ------------- ---- ---- -rw-r--r-- tictoc tictoc 4/28/2022 19:02 0 test (1).mp3 -rw-r--r-- tictoc tictoc 4/28/2022 19:02 0 test.mp3 -rw-r--r-- tictoc tictoc 4/28/2022 19:02 0 test1 (1).mp3 -rw-r--r-- tictoc tictoc 4/28/2022 19:02 0 test1.mp3 -rw-r--r-- tictoc tictoc 4/28/2022 19:02 0 test2 (1).mp3 -rw-r--r-- tictoc tictoc 4/28/2022 19:02 0 test2.mp3 -rw-r--r-- tictoc tictoc 4/28/2022 19:02 0 test3 (1).mp3 -rw-r--r-- tictoc tictoc 4/28/2022 19:02 0 test3.mp3 -rw-r--r-- tictoc tictoc 4/28/2022 19:02 0 test4 (1).mp3 -rw-r--r-- tictoc tictoc 4/28/2022 19:02 0 test4.mp3 -rw-r--r-- tictoc tictoc 4/28/2022 19:02 0 test5 (1).mp3 -rw-r--r-- tictoc tictoc 4/28/2022 19:02 0 test5.mp3
List all the files with "(1)" somewhere in the middle of the file name.
PS /home/tictoc/temp/testing/powershell> Get-ChildItem -Include *`(1`)* -Recurse Directory: /home/tictoc/temp/testing/powershell/music UnixMode User Group LastWriteTime Size Name -------- ---- ----- ------------- ---- ---- -rw-r--r-- tictoc tictoc 4/28/2022 17:37 0 test (1).mp3 -rw-r--r-- tictoc tictoc 4/28/2022 17:37 0 test1 (1).mp3 -rw-r--r-- tictoc tictoc 4/28/2022 17:39 0 test2 (1).mp3 -rw-r--r-- tictoc tictoc 4/28/2022 17:39 0 test3 (1).mp3 -rw-r--r-- tictoc tictoc 4/28/2022 17:39 0 test4 (1).mp3 -rw-r--r-- tictoc tictoc 4/28/2022 17:39 0 test5 (1).mp3 Directory: /home/tictoc/temp/testing/powershell/music2 UnixMode User Group LastWriteTime Size Name -------- ---- ----- ------------- ---- ---- -rw-r--r-- tictoc tictoc 4/28/2022 19:02 0 test (1).mp3 -rw-r--r-- tictoc tictoc 4/28/2022 19:02 0 test1 (1).mp3 -rw-r--r-- tictoc tictoc 4/28/2022 19:02 0 test2 (1).mp3 -rw-r--r-- tictoc tictoc 4/28/2022 19:02 0 test3 (1).mp3 -rw-r--r-- tictoc tictoc 4/28/2022 19:02 0 test4 (1).mp3 -rw-r--r-- tictoc tictoc 4/28/2022 19:02 0 test5 (1).mp3
Find the files and delete them:
PS /home/tictoc/temp/testing/powershell> Get-ChildItem -Include *`(1`)* -Recurse | RemoveItem
or the short and sweet one-liner with aliases:
PS /home/tictoc/temp/testing/powershell> gci -i *`(1`)* -r | ri
Final result:
PS /home/tictoc/temp/testing/powershell> Get-ChildItem -Recurse Directory: /home/tictoc/temp/testing/powershell UnixMode User Group LastWriteTime Size Name -------- ---- ----- ------------- ---- ---- drwxr-xr-x tictoc tictoc 4/28/2022 19:14 4096 music drwxr-xr-x tictoc tictoc 4/28/2022 19:14 4096 music2 Directory: /home/tictoc/temp/testing/powershell/music UnixMode User Group LastWriteTime Size Name -------- ---- ----- ------------- ---- ---- -rw-r--r-- tictoc tictoc 4/28/2022 17:33 0 test.mp3 -rw-r--r-- tictoc tictoc 4/28/2022 17:36 0 test1.mp3 -rw-r--r-- tictoc tictoc 4/28/2022 17:38 0 test2.mp3 -rw-r--r-- tictoc tictoc 4/28/2022 17:38 0 test3.mp3 -rw-r--r-- tictoc tictoc 4/28/2022 17:38 0 test4.mp3 -rw-r--r-- tictoc tictoc 4/28/2022 17:38 0 test5.mp3 Directory: /home/tictoc/temp/testing/powershell/music2 UnixMode User Group LastWriteTime Size Name -------- ---- ----- ------------- ---- ---- -rw-r--r-- tictoc tictoc 4/28/2022 19:02 0 test.mp3 -rw-r--r-- tictoc tictoc 4/28/2022 19:02 0 test1.mp3 -rw-r--r-- tictoc tictoc 4/28/2022 19:02 0 test2.mp3 -rw-r--r-- tictoc tictoc 4/28/2022 19:02 0 test3.mp3 -rw-r--r-- tictoc tictoc 4/28/2022 19:02 0 test4.mp3 -rw-r--r-- tictoc tictoc 4/28/2022 19:02 0 test5.mp3
Random aside, running PowerShell on Linux always feels like some whacky voodoo.
- 2
- 1
-
13 hours ago, Storm-Chaser said:
The maximum number of threads is 32 threads per slot. To get around this maximum, you just need to use miultiple CPU slots. I'll quote myself from the other thread:
QuoteF@H works with threads rather than cores, so if you wanted to max out the system I would probably use three CPU folding slots, two slots at 20 and one at 26 (assuming you have a pair of 18core 36 thread CPUs in that machine). That will leave three full cores free, two to feed the GPUs and one for additional system overhead.
- 3
- 2
-
3 hours ago, Avacado said:
I really appreciate that. But I'm going to pass. They need to more clearly label the products being sold. If it's LGA 3647 or LGA 3647-1, that needs to be evident. I know I can get a new intel 3647 for around $400 from ebay, but I want dual socket. If I can find a dual socket model, I might keep everything and wait a few years to hop on it. But I don't think a dual socket board exists for the Phi. Hey, at least I learned a bunch from this.
I don't belive any dual socket boards ever hit the consumer market. Most of the high density Phi clusters were/are running on blades, so 4 boards/processors per 2U rack space.
-
56 minutes ago, Storm-Chaser said:
Is folding@home able to take full advantage of the CPUs? How does that efficiency compare (GPU vs CPU)? reason being I do have a 24/7 rig I could fold with, dual procressor but weak GPU.
F@H will maximize CPU usage, but many of the simulations are highly parallelizable, so those calculations are more suited to GPUs. Once upon a time there were some very large tasks that were tailored to many core CPUs, but most of that work has been ported over to OpenMM to run on GPUs.
That being said, there are a number of projects that are CPU work only, and they are doing very interesting work. Here is a list of the currently active projects: https://apps.foldingathome.org/psummary The links in the project column will take you to a description of the project.
The OPENMM_22 core is GPU work, and GRO_A8 and GRO_A7 are CPU only. The CPU core has been optimized over the last 4-5 years, and now takes advantage of AVX instructions so it will put the CPU to work.
- 2
-
2 hours ago, Avacado said:
Kind of freaking out right now. I just read this
LGA 3647 - Wikipedia
EN.WIKIPEDIA.ORGApparently there are 2 versions of the P socket. But I have no way of knowing if the board that I bought is compatible with the Phi chips. Knowing my luck it wont be. I will find out tonight, but i'm kind of pissed that it's not easy to see.
X11DAi-N | Motherboards | Products | Super Micro Computer, Inc.
WWW.SUPERMICRO.COMThe board you linked (X11DAi-N) is socket P0 for Xeon Scalable-SP 1st and 2nd gen CPUs.
This is the only ATX form factor Supermicro board that I am aware of that supports Knight Landing. https://www.supermicro.com/products/motherboard/Xeon_Phi/K1SPE.cfm
There might be others, but I've never seen a working one in the wild other than the boards from the Supermicro and ASRock Rack developer workstations. I think Intel originally planned on Knight's Landing/Mill being compatible with all 3647 boards, but scrapped that plan when they EOL'd the Xeon Phi. I am pretty sure (but not positive) that you need a Xeon Phi specific board.
- 1
-
21 hours ago, Storm-Chaser said:
Okay I think everything is more or less sorted. @Avacadosorry about the temps in F!
The second CPU is slowly ramping up. The "CPU" portion is slowly creeping up in term s of CPU usage. I had not accounted for that in my last post.
So with fan settings on 3 out of 6 which is still pretty quiet I'm getting the following temps. But you are probably right once it fully ramps up CPU usage I will activate takeoff - go around power.
Right now cpus sitting at:
CPU0: 140*F
CPU1: 120*F
You can check core temp in the lower right corner and see I'm already pushing 290W
If you really want to put it through it's paces and fully utilize the CPU, I would add additionl CPU slots. F@H works with threads rather than cores, so if you wanted to max out the system I would probably use three CPU folding slots, two slots at 20 and one at 26 (assuming you have a pair of 18core 36 thread CPUs in that machine). That will leave three full cores free, two to feed the GPUs and one for additional system overhead.
The lastest F@H CPU core uses AVX, so expect a very heavy system load.
- 1
- 1
-
-
6 hours ago, Storm-Chaser said:
What am I looking at for an impact on performance with dual rank memory? I'm seeing 3-5% average improvement from a bit of research. In terms of bandwidth. Do two single rank modules work as like one dual rank module?
That is correct. Two single rank dimms per channel is effectively equal to one dual rank dimm per channel. With dual ranked dimms you would effectively be running quad rank if you populate all 8 slots, but generally speaking depending on the workload you will see a bump in performance.
I never really tuned my 2P v3 system, because I moved over to Threadripper shortly after acquiring it. The best performance on my Xeon v2 system, was using dual ranked dimms, but only installing 4 dimms per socket. My ASRock Rack board allowed me to bump the memory speed over the default for my sticks, so I was able to run 1333MHz sticks at 1866MHz. The board wasn't stable at 1866MHz with all 16 slots populated. Running in quad channel with 4 dimms per socket at the higher speed was the most performant setup, and I only moved away from that when I started to need more than 64GB of memory.
- 1
-
1 hour ago, Diffident said:
That's crazy. I've never broken blades off of a fan.
Don't worry about fan curves, servers are supposed to sound like jet engines.
The fan was spinning at 800 rpms when "someone" decided to put the front cover on the case, and might have inserted one of the cover standoffs directly into the path of the fan blades.
All the jet engines are going out to pasture, and I will basically end up with nearly the same overall performance and sweet whisper silence.
-
Octal rank is different than octal channel. Octal rank refers to the number of ranks per dimm. LRDIMMs generally come in quad rank and octal rank, which allows increased capacity per dimm. The increased capacity of LRDIMMs trades maximum capacity for performance.
In regards to whether or not it's octa channel, the platform itself (2 quad channel CPUs) technically has 8 channels, but each CPU can only address 4 channels locally. While the QPI is a fast path (8GT/s on Xeon v3) it's no where near as fast as local memory access. Generally speaking taking the performance hit is worth it due to the doubling of addressable memory.
Just saw your reply and that is all correct.
Specifically for your system the setup is good. The memory you have is single rank, so if you wanted to increase performance, you could look at getting some dual rank dimms which would bump up the memory throughput. For maximum performance anything above dual rank will generally be slower, and is only worth it if you actually need the additional capacity.
- 2
- 1
-
Actually it's still quad channel. Only Epyc, Threadripper Pro, and newer Xeon Gold/Platinum CPUs have 8-channel memory controllers.
Anything that is crossing the QPI to a far NUMA node, is going to suffer a hit in latency and performance. No idea how AIDA works, but in my experience, Windows generally tends to be less NUMA aware, so if the OS scheduler is not pinning processes to local memory then there will be a drop in performance.
-
After getting the 10 HDDs mounted, I decided to do some temperature testing to set the fan curve for the case fans. Had a derp moment while putting the front cover back on, and now I'll be installing a new front intake fan.
It might actually be a good thing, since I ordered a Silverstone AP183 which should push quite a bit more air than the Noctua.
I'll be cutting out all the unnecessary mounting points behind te fan which should help to increase airflow and reduce noise.
- 1
Amd and linux?
in Folding@Home
Posted
I've been running AMD GPUs on Linux since back in the bad old fglrx days, and outside of compute the current state of the drivers is better than it has ever been. @Bones your issue with the Radeon VII was probably due to the rather outdated kernel that ships with Mint. I don't know how much Mint backports to their kernel, but FYI for anyone that is running an AMD GPU, the AMD kernel driver receives constant updates and improvements and you will almost always be better served running a more up-to-date kernel than what many distros ship.
For compute on AMD, AMD is re-working a bunch of their docs around ROCm (as evidenced by the link being "ROCm_5.1.1" even though it points to ROCM 5.1.3 ). Here is a link to the current docs: https://docs.amd.com/category/ROCm_v5.1.1
@bonami2 Technically the 6800XT is not listed in AMD's docs as "officially" supported on the ROCm stack, but most everything I've tested with my 6900XT works including F@H. Once you get ROCm installed you should be good to go on your 6800XT. I can't confirm if the posted install method from Diff works. I run an "unsupported" distro, but what Diff posted, and the link to the GitHub issue, looks good to me so it's worth a shot.