Posts by ProDigit

21) Message boards : Questions and problems : Does the extra cache of the new 5800X3D increase crunching? (Message 108342)
Posted 8 Jun 2022 by ProDigit
Post:
I found a semi-suitable answer on the Folding forums.
At least for their client, a 5800x vs 5800X3D, are nearly identical in performance.
The X3D has about like 1% higher performance, despite it having a few hundred megahertz lower CPU speeds.
So the performance drop of the lower CPU frequency of the X3D, gets compensated in performance by the extra 3D RAM.
Quite literally, for folding the 5800x folds at 320k PPD, the 5800X3D folds at 322k PPD.

In most cases, it makes more sense buying a 5900x or 5950x, or even a 3900x, and disable HT, and you'll be running WUs much more efficient and fast.
Not only do you double the L-Cache per thread on the 3900/5900 series CPUs (as you cut the threads down by 50%), but also, the 5900x/5950x has 16 full cores, while the 5800x3D only has 8 full cores, 16 threads.
They say on average a dual threaded core roughly increases performance of a single core by just 33%.

The lack of L-Cache doesn't really seem a problem for most projects.
Disabling HT/SMT will allow the CPU to run cooler, and at higher frequencies, but does come with a drop in performance.
In this, the CPU follows the pattern of the GPU, where more cores generally is better, even if it's lacking L-Cache; and the CPU loses some cycles filling up that L-Cache between runs.
This appears to be the case all the way up to about 20 threads, where both L-Cache and believe it or not, the RAM speed, will become the limiting factor on crunching data.

Some tests have shown that the performance on a 3950x (16C/32T) seems to level out at around 20-22 threads.
Any more, and the CPU doesn't perform any faster (probably could also be power throttled).
However, once you hit closer to 28-30 threads, the CPU actually starts losing performance over less threads, to the point where 32 threads are no faster than 26 to 28 threads!
And that's something quite strange!
In the case of the 3950x, it may be better to test between disable or enable SMT, and not sure if that performance problem persists on the 5950x.

PPD wise, a 3900x gets 380kPPD, while the 5800X3D gets 322k PPD, or a PPD difference of 29.5%.

Considering that you can now get a much newer, faster 5900x for much cheaper in price than an older 3900x right now ($~380 vs $550), or the 5950x (16C/32T)for the price of a 3900x (12C24T), it makes only sense to go for a 5900x or 5950X if you want to do any CPU intensive computations.
22) Message boards : Questions and problems : Does the extra cache of the new 5800X3D increase crunching? (Message 108296)
Posted 1 Jun 2022 by ProDigit
Post:
AMD Ryzen 7 5800X3D, how much does it speed up data crunching on a variety of cpu project?
23) Message boards : Android : Android app can't be found in google search (Message 107969)
Posted 27 Apr 2022 by ProDigit
Post:
Funny!
I've been a Boinc user for so many years, and never used the download_all link (for my Android/Linux clients).
Didn't even know it existed lol!
Always used Gianfranco's versions
24) Message boards : Questions and problems : Raspberry Pi, BOINC Manager versions / Fresh install procedures on PiOS (Message 107941)
Posted 23 Apr 2022 by ProDigit
Post:
I'm still successfully running a lot of WUs on a 3 year old Linux and Boinc system.
Some projects won't run anymore, but I don't have the time to update every one of them to the latest, without running the danger of disabling some functions.
Like my recent upgrade disabled my Intel IGP WUs from the Collatz, while older versions of the OS still run with it.

The OS and the manager usually do very little to the WUs. The WUs themselves are made to run on most versions of Boinc and Linux, some even containing a docker style format, where the WU is ran in a virtual environment.
25) Message boards : Questions and problems : Loud fan but not overheating (Message 107940)
Posted 23 Apr 2022 by ProDigit
Post:
Simple,

1- Your bios fan profile is set up to ramp up the fan at certain temperatures.
Many fans run at 100% under 80% load, or when reaching 75+F.
Adjust your fan curve in Bios.

If you can't do that, the other option is that the fan is running high because the hot air is not properly exiting the case.
It's just floating around the cooler, basically making it suck in hot air.
What you can do to that is:
1- Create a duct from cardboard, paper, or semi-hard plastic plates, like thick foil.
Glue them with hot glue or tape, in a shape that will make the hot air exit the back of the case.

2- Install a case fan on the back part of the PC, sucking out the hot air in the back, alleviating the load on the CPU fan.


You can also swap the CPU fan for a curve bladed fan. They usually are quieter.
26) Message boards : GPUs : Water cooling (Message 107777)
Posted 9 Apr 2022 by ProDigit
Post:
The main problem with waterblocks is still insufficient cooling.
Most waterblocks only cool the top part of the GPU.
The bottom part also needs cooling.
They also sometimes only cool the GPU core itself, not the RAM.
The RAM needs at least a 80mm 12V fan running at 5V - kinda stream of air, over a heat sink solution.

It's always best to use a hybrid system.
Curve bladed 12V fans on 5V (usually case fan speed set to ~20%) make less noise than the water pump of a water cooled system.
They can also improve the cooling of passive cooled water systems, or even passive cooled GPUs, like GT 1030s or maybe even more future GPUs.
They use less than 0.5W to run, so it's a no-brainer to use.
At 12V they may use up to 2.5W, which is about 1% of a modern GPU's power draw.
27) Message boards : Projects : Cosmology@home error message (Message 107729)
Posted 5 Apr 2022 by ProDigit
Post:
Cosmology is online, but not enough work, according to BAM stats.
28) Message boards : Projects : Looking for memory-intensive projects (Message 107648)
Posted 30 Mar 2022 by ProDigit
Post:
Rosetta, MLC, LHC, Cosmology, Qchempedia, mind modeling....

Also helps if you have more cores (like, some projects like milkyway, can make use of as many cores as you have in a single WU).
29) Message boards : Projects : Rosetta@home needs help (Message 107647)
Posted 30 Mar 2022 by ProDigit
Post:
.... One thing that a number of projects have adopted is that posters must meet certain levels of activity on that project such as completing a number of tasks. ...

I always thought it would make sense to have a general BOINC forum, connected to a user account. Say, my name here is the same as my name on most projects.

If no automatic integration of user name, what prevents a user to post as one that doesn't have a forum profile, but has a high processed wu count?

The whole Boinc forums of each project should all come under 1 forum umbrella. Projects pay a small annual fee (like $100 or so) to keep the forum servers running, and in return users only need to make 1 user account for all past, current and future projects they may want to join.
30) Message boards : Android : No jobs for android? (Message 107646)
Posted 30 Mar 2022 by ProDigit
Post:
Thanks!
That worked!
Still strange that I didn't need to do this on the Galaxy A51.

Will add some lhc, Einstein, and other projects on my older TV box servers soon.
31) Message boards : Android : No jobs for android? (Message 107644)
Posted 30 Mar 2022 by ProDigit
Post:


The top one is the A72, the bottom one is A51.
32) Message boards : Android : No jobs for android? (Message 107643)
Posted 30 Mar 2022 by ProDigit
Post:
That sounds more like a comms problem than a CPU problem. Einstein is quite strong on https security. What does the event log say?


Client Messages

Mar 30, 2022 3:29:20 AM|Einstein@Home|File Arecibo_full.jpg exists already, skipping download

Mar 30, 2022 3:29:20 AM|Einstein@Home|File Android.jpg exists already, skipping download

Mar 30, 2022 3:28:42 AM|Einstein@Home|Finished download of Pulsars_crab_xr.jpg

Mar 30, 2022 3:28:42 AM||Internet access OK - project servers may be temporarily down.

Mar 30, 2022 3:28:41 AM|Einstein@Home|Finished download of Pulsars_schem2.jpg

Mar 30, 2022 3:28:41 AM|Einstein@Home|Finished download of Pulsars_schem1.jpg

Mar 30, 2022 3:28:41 AM|Einstein@Home|Finished download of Pulsars_crab_opt.jpg

Mar 30, 2022 3:28:41 AM|Einstein@Home|Finished download of Pulsars_Manhattan.jpg

Mar 30, 2022 3:28:41 AM|Einstein@Home|Finished download of Pulsars_J2007.jpg

Mar 30, 2022 3:28:41 AM|Einstein@Home|Finished download of Parkes_full.jpg

Mar 30, 2022 3:28:41 AM|Einstein@Home|Finished download of LIGO_vacuum.jpg

Mar 30, 2022 3:28:41 AM|Einstein@Home|Finished download of LIGO_seisisol.jpg

Mar 30, 2022 3:28:41 AM|Einstein@Home|Finished download of LIGO_schematic.jpg

Mar 30, 2022 3:28:41 AM|Einstein@Home|Finished download of LIGO_optics.jpg

Mar 30, 2022 3:28:41 AM|Einstein@Home|Finished download of LIGO_laser.jpg

Mar 30, 2022 3:28:41 AM|Einstein@Home|Finished download of LIGO_Livingston.jpg

Mar 30, 2022 3:28:41 AM|Einstein@Home|Finished download of LIGO_Hanford.jpg

Mar 30, 2022 3:28:41 AM|Einstein@Home|Finished download of GW_BBH2.jpg

Mar 30, 2022 3:28:41 AM|Einstein@Home|Finished download of GW_BBH1.jpg

Mar 30, 2022 3:28:41 AM|Einstein@Home|Finished download of Fermi_satellite.jpg

Mar 30, 2022 3:28:41 AM|Einstein@Home|Finished download of Fermi_grsky.jpg

Mar 30, 2022 3:28:41 AM|Einstein@Home|Finished download of Arecibo_platform.jpg

Mar 30, 2022 3:28:41 AM|Einstein@Home|Backing off 00:02:28 on download of Arecibo_full.jpg

Mar 30, 2022 3:28:41 AM|Einstein@Home|Temporarily failed download of Arecibo_full.jpg: transient HTTP error

Mar 30, 2022 3:28:41 AM|Einstein@Home|Backing off 00:02:47 on download of Android.jpg

Mar 30, 2022 3:28:41 AM|Einstein@Home|Temporarily failed download of Android.jpg: transient HTTP error

Mar 30, 2022 3:28:41 AM||Project communication failed: attempting access to reference site

Mar 30, 2022 3:28:40 AM|Einstein@Home|Finished download of Pulsars_vela.jpg

Mar 30, 2022 3:28:17 AM|Einstein@Home|Finished download of einstein_icon.png

Mar 30, 2022 3:28:15 AM|Einstein@Home|Started download of Pulsars_vela.jpg

Mar 30, 2022 3:28:15 AM|Einstein@Home|Started download of Pulsars_schem2.jpg

Mar 30, 2022 3:28:15 AM|Einstein@Home|Started download of Pulsars_schem1.jpg

Mar 30, 2022 3:28:15 AM|Einstein@Home|Started download of Pulsars_crab_xr.jpg

Mar 30, 2022 3:28:15 AM|Einstein@Home|Started download of Pulsars_crab_opt.jpg

Mar 30, 2022 3:28:15 AM|Einstein@Home|Started download of Pulsars_Manhattan.jpg

Mar 30, 2022 3:28:15 AM|Einstein@Home|Started download of Pulsars_J2007.jpg

Mar 30, 2022 3:28:15 AM|Einstein@Home|Started download of Parkes_full.jpg

Mar 30, 2022 3:28:15 AM|Einstein@Home|Started download of LIGO_vacuum.jpg

Mar 30, 2022 3:28:15 AM|Einstein@Home|Started download of LIGO_seisisol.jpg

Mar 30, 2022 3:28:15 AM|Einstein@Home|Started download of LIGO_schematic.jpg

Mar 30, 2022 3:28:15 AM|Einstein@Home|Started download of LIGO_optics.jpg

Mar 30, 2022 3:28:15 AM|Einstein@Home|Started download of LIGO_laser.jpg

Mar 30, 2022 3:28:15 AM|Einstein@Home|Started download of LIGO_Livingston.jpg

Mar 30, 2022 3:28:15 AM|Einstein@Home|Started download of LIGO_Hanford.jpg

Mar 30, 2022 3:28:15 AM|Einstein@Home|Started download of GW_BBH2.jpg

Mar 30, 2022 3:28:15 AM|Einstein@Home|Started download of GW_BBH1.jpg

Mar 30, 2022 3:28:15 AM|Einstein@Home|Started download of Fermi_satellite.jpg

Mar 30, 2022 3:28:15 AM|Einstein@Home|Started download of Fermi_grsky.jpg

Mar 30, 2022 3:28:15 AM|Einstein@Home|Started download of Arecibo_platform.jpg

Mar 30, 2022 3:28:15 AM|Einstein@Home|Started download of Arecibo_full.jpg

Mar 30, 2022 3:28:15 AM|Einstein@Home|Started download of Android.jpg
33) Message boards : Android : No jobs for android? (Message 107640)
Posted 30 Mar 2022 by ProDigit
Post:
So I've created a school profile, and Einstein seems to work on my Galaxy A51 (with Samsung Exynos 7 9611), however it doesn't seem to work on my Galaxy A72 (with Snapdragon 720G CPU).
Initially, Einstein said 'downloading 13 WUs', but they never downloaded and just disappeared.
Ever since I get no new WUs, and zero downloads on the A72.

My A51 seems to have about 4 WUs loaded, and so far it looks like they're processing, and already has accumulated close to 1000 credits.

The A72 is on the same account, just a different phone.
It doesn't seem to have the 'school' designation on it though.
When I look in my Einstein@home dashboard, only the A51 gets the 'school' designation.

Perhaps Einstein is CPU specific?
Or perhaps the designation somehow didn't take on the new phone?
34) Message boards : Android : No jobs for android? (Message 107557)
Posted 23 Mar 2022 by ProDigit
Post:
Universe is just back online today. Was still offline yesterday, due to an SSD issue, the project manager sent an update about it a few days ago.

As far as Einstein, I have no success running it.
Occasionally I get Einstein WUs, but I think just like with Yoyo, or Moo!, they just end up with computation errors.

Ah, I see,
Einstein was set up to do GPU projects only.
What Einstein project runs on ARM?
Following projects are disabled on Einstein, because I don't need to run them on my Intel x86 CPU:

Binary Radio Pulsar Search (Arecibo)
Gamma-ray pulsar binary search #1
Gamma-ray pulsar search #5
Gravitational Wave search O2 Multi-Directional
Gravitational Wave search O3 All-Sky



The only Application that Einstein has for Android is Binary Radio Pulsar Search (Arecibo) (BRP4). If you set your Android devices to be in a different location "Home or School" etc you can enable that application for only Android devices.


Where do I enable it for only android devices?
The android client doesn't have any configuration for this.

I set it to school, and then did:
DEFAULT SET >> Set used for new computers: >>YES

So far it hasn't downloaded any new content yet.
35) Message boards : GPUs : Specifications for NVidia RTX 30x0 range? (Message 107556)
Posted 23 Mar 2022 by ProDigit
Post:
...the GA102 (and above, but not the A100) benefit from both an increase from 64 to 128 cores per SM, and the ability to process two FP32 streams concurrently.

It was my understanding that the RTX 3060 has cuda cores (shaders) operating either at 32bit INT, OR 32bit FPP.
It's not 2 streams concurrently. It's either/or.
32bit INT works faster if you don't need to have as precise numbers. 32bit FPP is more precise, and a bit slower (and less hardware supports it).

It's funny, because it reminds me of my audio modelling days.
I'd run samples and effects, and there was this effect that used 32 bit INT reverb, and the reverb sounded a bit more metallicky.
Meanwhile 32 bit float, sounded like a 'perfect' reverb.
So the human ear was able to differentiate between the two, much like the eye can see the difference between 32bit INT (255 values of RGB), and 32bit float (255 values of RGB, and 64 bit per pixel shaded).
It's probably close to the ear and eye's maximum perceivable range of colors and sounds; which is why I never got why some digital stomp box pedals were sold with 24 bit reverbs. They sound like trash.
Things like 3D polygons run just fine on INT. Not sure if 32bit Float would work in a 3D game environment..
Anyway, but that's off topic.
36) Message boards : The Lounge : An undoubtedly very fascinating thread about GPU capabilities running multiple tasks (Message 107555)
Posted 23 Mar 2022 by ProDigit
Post:
I guess that’s why I have over a dozen high end discrete GPUs?

No one is saying that you should use laptop GPUs for crunching as your main system. Just that you can use the smaller chips to get an idea of performance that you might see at full scale. 2x performance uplift from previous gen to the Xe models is nothing to sneeze at. The whole idea is to be excited for intel’s upcoming DISCRETE GPUs based on observed performance from what the laptop chips are showing us.


I'm just happy that there's a third player in the market.
That means more supply, and lower prices.
And perhaps more and faster development, regardless if team Blue is behind right now.
I believe Intel has what it takes to be a significant player in the market.
Perhaps bringing GT 1030 performance in the beginning, but building up to RTX performance soon.
Scaling isn't the issue. Lithography nodes aren't the issue.
It's Intel's stringent quality checks, reliability, and their ability to improve upon last gen products through their know how of CPUs.
GPUs and CPUs are very much alike in the way they are built.
And I think Intel will take about 2 years to be where AMD is today (after 7 years of research).
37) Message boards : Promotion : An idea for Boinc in the Anthropocene. (Message 107554)
Posted 23 Mar 2022 by ProDigit
Post:
Yes, free country! Let it be like that!
I do wish that GPUs will become more available, and more powerful,for more efficient folding/GPU crunching.

And we're about to see the new Nvidia 4000 GPUs coming out!

With that, a lot of crypto miners will buy them up, and releasing their 3000-series GPUs on the second hand market.
RTX 3000 series GPUs are great for Boinc.
Can't wait for RTX 7000 to come out now!
38) Message boards : Promotion : Boinc in a box. (Message 107553)
Posted 23 Mar 2022 by ProDigit
Post:
Very few, if any, ARM systems have access to dGPUs, and if they do, there are even less projects that support them.

Most GPU WUs are written to be ran on an X86 CPU and a PCIE GPU; or on an X86 CPU or ARM CPU.
There are no ARM GPU projects I know of.

But even if there were,
Most X86/64 equivalent systems out there, use anywhere between 1 to 3 GPUs on a desktop (16-24PCIE lanes), and 1 to 6 GPUs on a Server (128PCIE lanes), with the exception of some exotic hardware being able to run 8-12 dGPUs.
Both X86/64 CPUs as well as ARM CPUs, are manufactured at 12, 10, 8, 7, 6, 5, and pretty soon 3nm.
The power difference between a 5nm x86/64 and a 5nm ARM CPU is about the same, both at idle and full load.
X86 uses more power when it has more PCIE lanes; which ARM doesn't have.
So you might as well install those solar panels, get that $70k Tesla battery pack, and run yourself a $200k Nvidia DGX A100 if you want to save the environment.

If not, you can probably save a good $300k on buying a simple Linux desktop pc with 3x RTX 3090 GPUs, for a good $7k, and do 1/3rd to 1/4th the work at half the power consumption.
39) Message boards : Android : No jobs for android? (Message 107522)
Posted 22 Mar 2022 by ProDigit
Post:
Universe is just back online today. Was still offline yesterday, due to an SSD issue, the project manager sent an update about it a few days ago.

As far as Einstein, I have no success running it.
Occasionally I get Einstein WUs, but I think just like with Yoyo, or Moo!, they just end up with computation errors.

Ah, I see,
Einstein was set up to do GPU projects only.
What Einstein project runs on ARM?
Following projects are disabled on Einstein, because I don't need to run them on my Intel x86 CPU:

Binary Radio Pulsar Search (Arecibo)
Gamma-ray pulsar binary search #1
Gamma-ray pulsar search #5
Gravitational Wave search O2 Multi-Directional
Gravitational Wave search O3 All-Sky
40) Message boards : Android : No jobs for android? (Message 107511)
Posted 21 Mar 2022 by ProDigit
Post:
Seems like Einstein and Rosetta still don't work on Android 10/11/12, and WCG and Universe@home are currently going through some maintenance.
Any other projects that work? LHC doesn't seem to give out any jobs.


Previous 20 · Next 20

Copyright © 2024 University of California.
Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.