One interesting side effect from the change in numbers required to Validate certain results- my Inconclusives have gone from around 3% to 17%

---

Grant
Darwin NT

I know the beta server as ran out of tasks, but can someone point me in the right direction on how to use the beta server?
I have a 5700XT and would like to help out.

---

I'm a Drunk

You join Beta by using the Tools tab in the Manager and choose the Add Project selection. Then in the Project URL selection box add this URL:

https://setiweb.ssl.berkeley.edu/beta/

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

Thank you.
Done, that was easy :-)

---

I'm a Drunk

Thank you.
Done, that was easy :-)

Great. And thank you for being a Beta tester. The sooner we can prove out that the 8.24 app solves the AMD problem with the Navi cards, the sooner the app can be brought to Main for the masses.

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

Thank you.
Done, that was easy :-)

Great. And thank you for being a Beta tester. The sooner we can prove out that the 8.24 app solves the AMD problem with the Navi cards, the sooner the app can be brought to Main for the masses.

and hopefully, all of the people who don't monitor forums or private messages will eventually update their drivers to the new version that is supposed to work. I think both are required for the fix to be effective (app+driver).

---

Seti@Home classic workunits: 29,492 CPU time: 134,419 hours

All I need now are Work Units.....

---

I'm a Drunk

Another 3 have joined the list.

aplrapid 7807183
Leigh Green 123169
Vytautas Liesis 173783

Cheers.

I think both are required for the fix to be effective (app+driver).

A new app wasn't needed at Einstein. Only the new driver necessary for the 5700XT cards to finally start reporting valid work when paired with a card other than another Navi.

Guessing that that Einstein app didn't implement that iffy compiler flag that Eric had to remove on our app.[/quote]

---

Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)

Of course, the day I can't be near the computer is the day the beta splitter freezes. I'm copying in new data and unfreezing it.

---

@SETIEric@qoto.org (Mastodon)

[quote]Of course, the day I can't be near the compu
. . We all have to learn to live with Murphy's Law ... no matter how much it sucks :(

Stephen

<shrug>

The warning is still spreading:
https://www.hardwaretimes.com/opencl-compute-broken-on-amds-navi-cards-since-launch-all-radeon-rx-5700-5700-xt-gpus-affected/
What’s worse is that many Databases aren’t able to detect these invalid results, leading to possible corruption of OpenCL based Databases. In case you’re running an RX 5700 series card, kindly refrain from using it for OpenCL Compute. This includes Online Science Databases such as SETI, Geekbench, etc. We’ll let you know as soon as AMD fixes this issue.

https://www.reddit.com/r/Amd/comments/eblv26/psa_please_remove_your_amd_rx5700xt_from_setihome/
PSA: Please remove your AMD RX5700/XT from SETI@home now.

This issue has been ongoing since the RX5700 and RX5700XT was released. OpenCL compute is broken on these GPUs (unclear at this time if it’s hardware or drivers related) and they produce invalid results.

The problem is that these RX5700s are cross validating their incorrect results with each other on occasion. If left unchecked this has serious implications for the integrity of the science database.

https://www.pcgamesn.com/amd/radeon-rx-5700-xt-seti-home-aliens
AMD’s RX 5700-series graphics cards are turning up incorrect results in SETI@home, a crowd-supported “volunteer computing” experiment searching the vast emptiness of space for signs of extraterrestrial intelligence.

https://www.eteknix.com/seti-may-ban-amd-5700-gpus-from-searching-for-aliens/
In a report via TechPowerUp, it seems that the community is currently debating whether owners of AMD 5700 graphics cards can participate.

---

I was the one who made that PSA post on reddit. I knew it got good visibility on the AMD sub, but I had no idea that people wrote articles about it.

---

Seti@Home classic workunits: 29,492 CPU time: 134,419 hours

I was the one who made that PSA post on reddit. I knew it got good visibility on the AMD sub, but I had no idea that people wrote articles about it.

That is good that the warning spread, but once beta proves it works, articles need to be posted to say it gas been fixed!

---

Seti@home classic: 1,456 results, 1.613 years CPU time

I was looking at StrayCat’s results over at beta, to get an idea of relative performance.

Judging based on the WU times compared to the cards it’s being stacked up against on the validates tasks, it looks similar in performance to a GTX 1060 or so. That seems a bit disappointing, I think it should be about 2x that.

Can anyone link to other 5700 systems at beta so I can see if others are having the same level of performance, or if it’s just his system.

StrayCat, can you comment on GPU load percentage? I wonder if since BOINC is only reporting half the CU count, that it’s only using half of the card or something. Or maybe it’s the removal of the compiler parameter or the driver itself that’s hampering performance.

---

Seti@Home classic workunits: 29,492 CPU time: 134,419 hours

The performance hit may be real, and a direct result of Eric's comment about having had to turn off one of the optimization flags.

---

Bob Smith
Member of Seti PIPPS (Pluto is a Planet Protest Society)
Somewhere in the (un)known Universe?

So far, so good:

1219 results have been through the validator
1158 (95%) have validated
35 (2.9%) are inconclusive and may validate later
26 (2.1%) are errors, but don't appear to be computation errors.

If things continue this way I will probably put the new app versions on the main projects tomorrow or tuesday.

---

@SETIEric@qoto.org (Mastodon)

If you compare the CPU times you will see StrayCat's are Very Low, https://setiweb.ssl.berkeley.edu/beta/results.php?hostid=89082&offset=20
It appears StrayCat has his CPU setting at 100%. Others have a Much Higher CPU Time, and Much better Run times, https://setiweb.ssl.berkeley.edu/beta/results.php?hostid=89092
I'd say 90% of the people at SETI are Strangling their GPUs by Not reducing their CPU settings to below 100% when running GPU tasks.
It certainly doesn't help by having some people going around telling users to set their CPU setting to 100%...and then make an app_config to lower the number of tasks.
What usually happens is the User can do the first part, raise the CPU setting to 100%, BUT, fails to make the app_config because it's more difficult. That leaves the GPU being Strangled for CPU time, and producing poor results.

Okay, running my RX 5700 XT at Beta. Got an AP.
Keep an eye on it via https://setiweb.ssl.berkeley.edu/beta/results.php?hostid=89104
(freed one core for the GPU)

Edit: an AP in 9m37s? Wow.
Second AP in 9m09s.
Third AP in 9m01s.

---

“Wow” fast? Or “Wow” Slow?

My 2070s routinely run them about that fast (sometimes faster, sometimes slower), when running 2 WUs at a time.

---

Seti@Home classic workunits: 29,492 CPU time: 134,419 hours