Noting this limitation appears to be specifically for Windows versions then what about Linux version code?
jsm

I'm not seeing anything specific to any operating system. The code applies to every host contacting the scheduler as far as I can tell.

---

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

A proud member of the OFA (Old Farts Association)

This is in the server code and as far as we are concerned is operating system independent, but a hard limit we, as users, can't get round.

---

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

but a hard limit we, as users, can't get round.

Don`t be so sure about that,

---

Look again at the code snippet which started this hare running:

const int MAX_NCPUS = 64;
    // max multiplier for daily_result_quota.
    // need to change as multicore processors expand

Used for daily_result_quota - nothing to do with 'in progress' limits. There's more:

int _max_results_day;
// MRD is dynamically adjusted to limit work sent to bad hosts.
// The maximum # of results sent per day is
// max_results_day * (NCPUS + NCUDA * cuda_multiplier).
// 0 means uninitialized; set to config.daily_result_quota
// -1 means this host is blacklisted - don't return results
// or accept results or trickles; just send it an error message
// Otherwise it lies in the range 0 .. config.daily_result_quota
// DEPRECATED: only use is -1 means host is blacklisted

(from https://github.com/BOINC/boinc/blob/master/db/boinc_db_types.h#L347)

but a hard limit we, as users, can't get round.

Run multiple clients.

int MAX_NCPUS

// DEPRECATED: only use is -1 means host is blacklisted

Which as we know . . . . <sarcasm_hat_on>works so well</sarcasm_hat_off>

Interested in the variable name as it is assumed it is also used to limit max tasks in progress per device. But I haven't found it elsewhere in the server code.

---

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

A proud member of the OFA (Old Farts Association)

int MAX_NCPUS

// DEPRECATED: only use is -1 means host is blacklisted

Which as we know . . . . <sarcasm_hat_on>works so well</sarcasm_hat_off>

The -1 value isn't generated automatically - it has to be set manually by staff. It's been used (in the past) by projects like CPDN, where the months-long tasks are worth protecting against rogue hosts.

int MAX_NCPUS

// DEPRECATED: only use is -1 means host is blacklisted

Which as we know . . . . <sarcasm_hat_on>works so well</sarcasm_hat_off>

The -1 value isn't generated automatically - it has to be set manually by staff. It's been used (in the past) by projects like CPDN, where the months-long tasks are worth protecting against rogue hosts.

Oh. . . . I was under the impression that was the mechanism to restrict sending work to bad hosts that can't return valid work ever and was automatic (supposedly).

---

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

A proud member of the OFA (Old Farts Association)

No, it's set manually in the BOINC database.

Oh. . . . I was under the impression that was the mechanism to restrict sending work to bad hosts that can't return valid work ever and was automatic (supposedly).

There is such a mechanism - badly implemented - but this isn't it.

Tom's Hardware redid the SPEC3 Workstation benchmarks with the new version of the benchmark that properly spans the two processor groups that the Windows OS breaks the processor into. Now the processor is showing the expected scaling when it can use all 128 threads across all the tests of the new benchmark.
https://www.tomshardware.com/features/amd_threadripper_3990x-spec-workstation_3-performance-update
Benchmarks

---

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

A proud member of the OFA (Old Farts Association)

Phew! That was close. I was placing my order for a 3990X bundle when the news came through that Seti@home was 'hibernating'. Having quickly checked that the date wasn't April 1st, I cancelled and am some £5000 better off.
I hope work will be available up to the end of the month as I am close to the top 1000 personally and my team to the top 500 and I would like to achieve those targets.
Its a shame really because the AMD push for ever more threads is a catalyst for high end arithmetical crunching.
jsm

Phew! That was close. I was placing my order for a 3990X bundle when the news came through that Seti@home was 'hibernating'. Having quickly checked that the date wasn't April 1st, I cancelled and am some £5000 better off.
I hope work will be available up to the end of the month as I am close to the top 1000 personally and my team to the top 500 and I would like to achieve those targets.
Its a shame really because the AMD push for ever more threads is a catalyst for high end arithmetical crunching.
jsm

I can tell a Seti@Home core junky when I see one ;)

I am very pleased with the AMD 3900x I had and with the 3950x upgrade I am running now. The thread count takes me back up into the server class count where I first experienced it but with significantly faster and more efficient processing. And it didn't cost as much as my Threadripper2 did.

And they help with my "core junky" fix.

Tom M

---

A proud member of the OFA (Old Farts Association).

Apparently things were a little slow. So I got a notice (aka commercial) on my Feed about a "huge" reduction in AMD 3900x prices.... in the small print, it came in a decrease in $30-$40 USD. I agree that is a decrease.

But come on now "Huge"?

I don't think so.

Tom

---

A proud member of the OFA (Old Farts Association).

Phew! That was close. I was placing my order for a 3990X bundle when the news came through that Seti@home was 'hibernating'. Having quickly checked that the date wasn't April 1st, I cancelled and am some £5000 better off.
I hope work will be available up to the end of the month as I am close to the top 1000 personally and my team to the top 500 and I would like to achieve those targets.
Its a shame really because the AMD push for ever more threads is a catalyst for high end arithmetical crunching.
jsm

Now you can join the I run nearly all gpu club..... you have all that money left over burning an asteroid in your pocket... So buy a raft RTX 2060 Supers and go crazy.

Tom

---

A proud member of the OFA (Old Farts Association).

Nothing beats the power savings of just running a host with a lot of cpu threads and some dinky graphics card to display the desktop. The power vultures are the gpus.

---

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

A proud member of the OFA (Old Farts Association)

Just for fun and with only a few days left I am going to move from the 'sweet spot' of threads on the 2990wx to 100% so that I can enjoy the sight of 64 simultaneous tasks on one cpu.
jsm

Just for fun and with only a few days left I am going to move from the 'sweet spot' of threads on the 2990wx to 100% so that I can enjoy the sight of 64 simultaneous tasks on one cpu.
jsm

And unless you have a really large monitor, having to scroll up/down to see all those threads.

More thread ripping to you!

You know, if you didn't care about the production you could run all 64 threads all the time ;)

Tom

---

A proud member of the OFA (Old Farts Association).

And I do have a large monitor but its still two full screens. Fascinating. But 'no tasks available' is the constant message.
jsm

And I do have a large monitor but its still two full screens. Fascinating. But 'no tasks available' is the constant message.
jsm

Sigh. If you are big into biology that is socially conscious you could try World Community Grid in their "maximum processing" mode. They are also running a Weather forecasting app for farmers in Africa. That takes up to 14 hours on my 3950x so it should keep a 2990wx threadripper busy forever... :)

I know that won't help your lust to see near 64 Seti@Home threads...... but it might make you nostalgic....

Tom

---

A proud member of the OFA (Old Farts Association).