You obviously;y do not understand the architecture of the 8+20 core processor in the OP's computer - it is VERY different to that which you and I are used to with our AMD Ryzens, and the "best" answer will be one that works with the heterogenous core speeds (8 run at about double that of the other 20).
So, first guess would be running multiple 4-core tasks, in the hope that 2 tasks would be run by the fast "performance" cores and the rest by the much slower "economy" cores. So, starting point would be 3 off 4-core tasks total, leaving some spare cores for the operating system and other uses. Once the OP has some idea of the runtime ranges at that level, and assuming there are no nasty (unexpected) interactions increase the number of concurrent tasks, see what happens to both timings and tasks per day.


As for modding client_state.xml
Risky, needs care and that is why I pointed the OP to using the simpler and more robust app_config.xml It is worth knowing that if an app_config.xml is found it over rules the "generic/global" client_state.xml file.
Actually client_state.xml can be modified, but great care has to be taken - large sections can be removed or changed without breaking it.

---

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

You obviously;y do not understand the architecture of the 8+20 core processor in the OP's computer
(...)
(8 run at about double that of the other 20)
(...)
So, starting point would be 3 off 4-core tasks total, leaving some spare cores for the operating system and other uses.

And you obviously not only don't understand what I have written, but also what was the question: currently Milkyway is using 16 of the 28 available threads and the OP wants to use more/all threads, not less.

And that's an 8C+12c processor with total 28 threads, not 8C+20c. The 8 P-cores have HT, the E-cores don't, that makes the speed difference smaller. But yes, I agree that 4-thread tasks should work for it too and are fully compatible with the other computers running Milkyway under this account, so maybe best choice in this case.

As for modding client_state.xml
Risky, needs care and that is why I pointed the OP to using the simpler and more robust app_config.xml

They should not modify anything, they should check wether the changes on the website were sent to the client. That was a pretty simple question I asked and now I'm waiting for the answer. That's how far we have come until now and we don't need to confuse people asking for help with more technical stuff than necessary.

---

They should not modify anything, they should check wether the changes on the website were sent to the client. That was a pretty simple question I asked and now I'm waiting for the answer. That's how far we have come until now and we don't need to confuse people asking for help with more technical stuff than necessary.

Then WHY did you even point them to that file in the first place? You totally confuse everybody with incorrect information based on your own inadequate understand of what is happening.

---

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

My message was a mistake but I couldn't cancel it.

And you obviously not only don't understand what I have written, but also what was the question: currently Milkyway is using 16 of the 28 available threads and the OP wants to use more/all threads, not less.

It will be worth waiting for the current storm of tasks that are ending in failures before starting any proper tuning of the OP's configuration.

---

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

They should not modify anything, they should check wether the changes on the website were sent to the client. That was a pretty simple question I asked and now I'm waiting for the answer. That's how far we have come until now and we don't need to confuse people asking for help with more technical stuff than necessary.

Then WHY did you even point them to that file in the first place?

How hard is it for you to read and understand what you quote? You quoted the answer on your question!

---

No, it is you does not understand that what you write is open to misinterpretation.

---

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

Yes, I saved the settings, which shows 7 on the web site. And app_version still shows 16.

Yes, I saved the settings, which shows 7 on the web site. And app_version still shows 16.

That's strange. In the MilkyWay@home preferences, are you using there only the "Primary (default) preferences"? And if yes, is the location for all your computers empty, i.e. it is not home/school/work?

BTW., after checking your other computers I noticed, that 7 threads per task isn't optimal for them. Unless you want to move the new computer to a different "location" and set 7 threads per task for it and leave the other at the default value, I suggest 4 threads per task, that will work for all your computers. With 7 threads per task you would have 1 empty thread on the two computers with 4 cores / 8 threads.

---

Even stranger is looking at the ratio of "clock" time to CPU time.
For task https://milkyway.cs.rpi.edu/milkyway/result.php?resultid=1034205122
There's a clock time of ~25minutes, and a CPU time of ~2.5hrs
This gives a CPU/clock of ~6, but the stderr says it ran on 16 "cpus" - that CPU is not behaving in a very good manner. Just a passing thought, is this the Intel load sharing kicking in and moving the job between the -P & -E cores during the run?
If this is the case it's going to be areal pain to work out what is the best number of threads to use for maximum throughput (or indeed a consistent throughput)



[edit to add]
I just looked at another random task, and it is showing a very similar pattern:
https://milkyway.cs.rpi.edu/milkyway/result.php?resultid=1034194027
Clock time ~52 minutes
CPU time ~5hrs 20 minutes
CPU/clock ~6
No record of number of threads/cores in use.....

---

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

As a comparison I've just looked at a couple of recent tasks from my computer
https://milkyway.cs.rpi.edu/milkyway/result.php?resultid=1034320770
Clock ~12.5 mins
CPU ~59 mins
6 of 16 cores used
CPU/clock ~4.7

https://milkyway.cs.rpi.edu/milkyway/result.php?resultid=1034312432
Clock ~3hrs 30mins
CPU ~16hrs 7mins
6 of 16 cores
CPU/clock
CPU/clock ~4.6

---

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

I don't see an option called "Primary (default) preferences." I captured an image of the MilkyWay@home preferences page but I couldn't figure out how to show it here. So here it is.

Resource share 130
Use CPU
Use AMD GPU
Use NVIDIA GPU
Is it OK for MilkyWay@home and your team (if any) to [img]email [/img]you?
Should MilkyWay@home show your computers on its web site?
Default computer location ---
Color scheme for graphics Tahiti Sunset
Maximum CPU % for graphics
0...100 20
Run only the selected applications (all applications)
If no work for selected applications is available, accept work from other applications? yes
Max # of simultaneous MilkyWay@home tasks No limit
Max # of threads for each MilkyWay@home task 7

Use CPU is checked, the two GPUs are not. email and web site permissions are checked.

And I am only using the top two computers in the "your computers" list, DESKTOP-5RMKDF7 and Lat7400-FYRQYY2. They are the two Windows 11 computers that I moved the other two to. Yes, the column "Location" is empty. FWIW, currently the laptop is running 8 WCG work units and using 90+% of the CPUs.

I am very grateful you are helping me with this, thanks.

Even stranger is looking at the ratio of "clock" time to CPU time.
For task https://milkyway.cs.rpi.edu/milkyway/result.php?resultid=1034205122
There's a clock time of ~25minutes, and a CPU time of ~2.5hrs
This gives a CPU/clock of ~6, but the stderr says it ran on 16 "cpus" - that CPU is not behaving in a very good manner. Just a passing thought, is this the Intel load sharing kicking in and moving the job between the -P & -E cores during the run?

Yes. This is likely Windows running all project applications on E-cores only because of their low priority. As I already posted somewhere above, the priority needs to be raised as described in this post, there was already another user with this issue.

---

I don't see an option called "Primary (default) preferences." I captured an image of the MilkyWay@home preferences page but I couldn't figure out how to show it here.

Well, it's not an option, it's a set of options. It's written on the blue line directly under "MilkyWay@home preferences". Do you have 3 buttons on the bottom of the page called "Add separate preferences for home", "Add separate preferences for school" and "Add separate preferences for work"? If yes, than you are using only the default preferences. Or check this view, there you have the blue line with "Project specific settings", Default, Home, School and Work and if under Home, School and Work there are only --- or -- and an "Add" in the bottom, than you don't have any additional settings for those locations.

Max # of threads for each MilkyWay@home task 7

Change that to 4 now, save it and see if than the server finally sends it to your computer. No idea, why it has not done that yet.

currently the laptop is running 8 WCG work units and using 90+% of the CPUs.

Than it's likely running at the thermal limit and throttling, should be pretty constant at 99%. Check temperatures with HWiNFO64. If they are at or close to 100°C, than this isn't healthy in the long term, so you should consider underclocking and eventually undervolting it, depending on how much experience you have. Underclocking is easy and can be done even from Windows energy settings. Well, at least it was possible in Win10, hopefully Microsoft didn't remove it in Win11. Or maybe it's at least possible to disable the turbo in the UEFI.

---

Don't see the blue line. However,there are three blue rectangles at the bottom: add preferences for...
So I guess I'm in Primary (default) preferences.

Changed Max # of threads for each MilkyWay@home task to 4. Updated. app_version still shows 16

Laptop temperatures nowhere near 100C, thanks.

It can take a good few hours for changes to propagate through the server (and your computer) for changes to take effect, but it is a bit random.


[edit to add]
Your laptop staying below 100C is good news.
Many newer laptops are very good at controlling their internal temperature by varying the speed of components, including chip clocks, fans, memory.

---

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

Don't see the blue line. However,there are three blue rectangles at the bottom: add preferences for...
So I guess I'm in Primary (default) preferences.

Yes, if there are three blue buttons, than you have only the default preferences.

Changed Max # of threads for each MilkyWay@home task to 4. Updated. app_version still shows 16

Well, if that doesn't work, we will need to create an app_config.xml file... but than I have to ask first: do you perhaps already have that file in your Milkyway project folder (usually C:\ProgramData\BOINC\projects\milkyway.cs.rpi.edu_milkyway on Windows)? This would explain, why the client ignores the settings on the website.


Anyway, first things first, your computer must crunch on all cores, in particullar on the P-cores, currently it seems it's only crunching on the E-cores. So you need to create a file called cc_config.xml in your BOINC data dir (usually C:\ProgramData\BOINC\) if you don't already have one in there and insert the following into it (using Editor or other plain text editor, not Word or something like that):

<cc_config>
 <options>
  <process_priority>2</process_priority>
 </options>
</cc_config>

If you alredy have that file, you only need to add the process_priority line in the options section.

---

OK, added cc_config.xml. Closed BOINC manager, restarted.
Event Log now shows using MilkyWay@home preferences.
I guess we should wait a while to see if 4 instead of 16 shows up in app_version.

Is it using now all cores? Check either with Task Manager or, even better, with HWiNFO64, as that shows which are P-cores and which are E-cores.

OK, we wait. You can also set Milkyway to no new tasks, let it finish all the tasks you have in cache, than remove it and add it again to get a clean start with the new settings if waiting won't help.

---

OK, I'm impatient, so I aborted all but the currently running WU, let it finish. Set all projects to no new tasks. Removed MilkyWay and reinstalled it. Set MilkyWay to new tasks, then Rosetta, WCG, and Einstein in that order.

Now it's running 7 MilkyWay WUs at 4 CPUs each. HWiNFO64 shows 8 P-cores and 12 E-cores. Task Manager briefly showed 85 - 90% CPU (when it was also running 3 Einstein WUs). Then it was 40-50. I don't know why it set the Einstein WUs to "waiting to run." Now it's back to 80-90. Keeping my fingers crossed.

Thanks much.