This is a different twist on the CPU hogging questions. When BOINC/SETI is running, it drives high CPU usage of the System.exe task in my WinXP/SP2 system. Since System.exe is a system task, high CPU does impact other processes and hurt user response time.

Example: With SETI suspended, Task Manager reports CPU 97% idle...when it is running, System.exe takes 55-70% of the CPU, and SETI gets the rest.

This problem appeared a few months ago...before that BOINC/SETI was getting almost all the idle CPU (as it should).

---

System.exe is a trojan. It has nothing to do with BOINC or SETI, nor is it a system process.

---

System.exe is a trojan. It has nothing to do with BOINC or SETI, nor is it a system process.

My My that is a pretty stupid answer.................
I have that same exact problem. Boinc v 5.4.9....and no system.exe isnt
a trojan..its a the Windows System process that controls who your system works
So far no one seems to know why it is doing this but I suspect it has
something to do with the Deferred Procedure Calls that may have been
changed in v5.4.9 (the way it communicates with the client...)

---

System is part of the Windows Idle checking routine. This process is a single thread running on each processor, which has the sole task of accounting for processor time when the system isn't processing other threads.

System Idle Process shows the actual output of System.

system.exe is a trojan.

---

My My that is a pretty stupid answer.................
I have that same exact problem. Boinc v 5.4.9....and no system.exe isnt
a trojan..its a the Windows System process that controls who your system works
So far no one seems to know why it is doing this but I suspect it has
something to do with the Deferred Procedure Calls that may have been
changed in v5.4.9 (the way it communicates with the client...)

You really need to do research before you challenge someone's answer. Check out Ageless's link that will direct you in the right way. System.exe is a trojan. Guaranteed or your money back. Run an Adware/spyware/antivirus program with updated defs if you don't believe me. The one thing I specialize in is x86 hardware and Windows' OS (all variations since v3.0 and the 8086). I guarantee you, if I am in doubt, I will say as much in my posts. If I say it as fact, it is fact and verifiable.

"Don't **** with a Jedi Master, son" - Mark Hamill on Jay and Silent Bob Strike Back.

---

Correction...the process using the CPU is (Image Name) System (not System.exe). Have run repeated virus and trojan scans, and this happens on a very fresh XP install.
So nonetheless, using Ageless' definitions, System is using over 50% of the CPU while BOINC/SETI is running.

---

Correction...the process using the CPU is (Image Name) System (not System.exe). Have run repeated virus and trojan scans, and this happens on a very fresh XP install.
So nonetheless, using Ageless' definitions, System is using over 50% of the CPU while BOINC/SETI is running.

That's a different story. Do you have a dual CPU, dual core, or HT enabled system? If so, you are seeing 50% because one of your "CPUs" isn't being utilized.

Please check here for information on how to make BOINC use both CPUs/cores/virtual CPUs.

---

The answer is no. I have a 2.8G P4, albeit a special one with a 400MHz front-side bus. Intel didn't make very many of these, but it is single-core. Upgrade CPU for a Dell 4300.

I had SETI doing 97+% on this config for a while last year...then it slowed down to the current lower level of performance. Mithotar might be on to something.

---

The answer is no. I have a 2.8G P4, albeit a special one with a 400MHz front-side bus. Intel didn't make very many of these, but it is single-core. Upgrade CPU for a Dell 4300.

I had SETI doing 97+% on this config for a while last year...then it slowed down to the current lower level of performance. Mithotar might be on to something.

Doubtfully.

If SETI isn't getting full 100% during idle times, then something else is taking the CPU's time away from it via a higher priority. By the way, even a fresh install of XP can quite easily get a virus/trojan within minutes of going online (I've seen it happen several times).

If SETI is only getting 97+% or less and the System process is actually using CPU cycles during idle times, then it has to be a different piece of software installed that's using the system process and thus taking away time from SETI, such as a web browser windows that's been open for a few hours/days and has loaded several instances of Flash or Windows Media player - it seems to take CPU cycles even when you're not doing anything, simply so the program can keep it's data updated. Of course, this was just an example. There are other possibilities.

Logically, if it was the way BOINC/SETI was programmed (i.e. a bug or some other anomoly), then it would happen on all machines, not just a few (i.e. it would happen on all systems using the v5.4.x client). Being that every one of my machines has the latest client installed and use the full 100% CPU process, I'm lead to believe there's something else about your configuration that's causing this. There's not enough logical evidence to support the theory that it's a change in BOINC's code.

---

By the way, even a fresh install of XP can quite easily get a virus/trojan within minutes of going online (I've seen it happen several times).

When worms were all the rage, it generally took less than 30 seconds for an unpatched system to become infected once connected to the net.

---

Grant
Darwin NT

I suspect it has
something to do with the Deferred Procedure Calls that may have been
changed in v5.4.9 (the way it communicates with the client...)

LOL so any application can execute its code under System process using DPC? Try to install boinc as a service so it can't commucate with client and see whether problem remains. I think, only drivers can hang System process such as firewalls, antiviruses, virtual drives. BOINC is just unable to do this.

Sorry about the tone Ozz but your reply was a bit "off the mark"
Besides you arent the only one in the world who has worked on computer
hardware since the 8088 era <grin>. I now slog away for a big company as a FE
for server(s) and networks.

The problem is that ADAware - SpyBot - Rootkit Revealer etc..show no problems
at all (and each gets run on a weekly basis). Start BOINC and within 10 min the System (and DPC process shoot up into the 50% plus range) stop BOINC and they
both drop within seconds. At no other time than when BOINC is active does this
occur. I have tried BOINC as a service with the same exact result.

Using a variety of tools I have narrowed the problem to this :
whatever is going on has to do with this .....APCI.sys+0xd8c8...somthing in BOINC is triggering this on some PCs (chipset or steping code of a CPU possibly)
but this is quite real and in no way related to a virus or trojan.

I have 4 other P4 systems running BOINC without a hitch .....just this particular one (which ran Seti@home for years and ran BOINC without a hiccup until v5.4.9...
Something is afoot in the BOINC code.........

I'll keep digging as the goal here is to identify and hopefully correct this
so it doesnt take 60+ hours to do a WU......

---

That's cool Mithotar. Believe me, I know I'm not the only one that's been in computers forever and a day. I was just being facetious with my harsh reply (part of my dry sense of humor), but I am serious in that System.EXE is a trojan. The System process and System Idle Process are not, but System.EXE most definately is. I was just being very specific to file names as that is what is posted. How am I to know that he didn't really mean System.EXE?

Plus, I have BOINC v5.4.9 running on all of my machines without this issue. My machines are sort of a museum with varying hardware and Windows software ranging from (at least what will run BOINC) an old Pentium 233MMX and a IDT Winchip 180MMX with an old 430TX chipset and a COMPAQ proprietary chipset, to a Cyrix MII 366MHz on a VIA MVP3 chipset (also a K6-2 450 and 500 on the same chipset), to a Pentium II on a 440LX, to a PIII 1GHz on a 440BX chipset and a P4 3GHz on a 875P chipset (plus a few others).

My question is, if it was really the code, why wouldn't it happen on all platforms? I cannot accept it as a chipset or stepping code error because BOINC does not function on that level. If it worked more closely with hardware, I'd accept that answer, but it doesn't.

...but this is quite real and in no way related to a virus or trojan.

Perhaps, but the name of the post was System.exe, which most definately is a virus. I'm not trying to nitpick, but being accurate makes all the difference. Since there is a difference between the System process and System.exe, I assumed he scanned his harddrive and found a file called System.exe and I wanted to let him know that it was a trojan. Had he simply said "System", I would have answered a little differently, but I still have to believe it's particular to his/your system and not the BOINC code. ***But PLEASE understand that I am not saying it isn't possible - just that it doesn't appear that way with the offered theories.***

---

whatever is going on has to do with this .....APCI.sys+0xd8c8...somthing in BOINC is triggering this on some PCs

I've done some digging and have a potential solution. ACPI.SYS (as it's supposed to be, not APCI.SYS) is the Advanced Configuration and Power Interface driver, which controls all BIOS and System level triggers for hardware related heat/power issues.

My theory is that since BOINC runs the system at 100% (full CPU usage), in a poorly ventilated system this might trigger the overheat driver in ACPI.SYS causing the system to drive the CPU down to slower levels so as to cool down the system (as set in the BIOS; CPU will typically run 25-50% or 50-75% depending on BIOS setting).

The only hole in my theory is that it should have been triggered with older BOINC versions as this is built-in to the OS's ACPI support. But it might still be worth looking into.

---

I finally got this problem solved.......with Ozz's suggestions I started to
look into the cooling issue........and there was the problem.

The PC was running @ 65-67C when Boinc wasnt running ..5 minutes after BOINC
started the temp went up to 72-74C....but the main culprit was either the fan
or the system board because as the the temp went up the fan slowed down. Most of the time the fan virtually stopped when the temp hit 74C.....very odd behavior.

The solution was both a new fan and a new systemboard ........now the system runs
at 52C normally and 55-58C when BOINC cranks up.

All is good now and so I am back to having my "ranch" of 3 x P4 2.0s and 2 x Dual
1.7G Xeons cranking again.

Kudos to Mr Ozz....

---

OzzFan may indeed have the solution...yet to verify, but the theory certainly is possible in my "upgraded" CPU and cooling solution. This never happened in the stock system...and the slow down does occur a few minutes after BOINC starts up.

Mea culpa on the post title, but in my defense, I corrected with another post 4 days later...it would be a nice feature to let the original writer edit the thread title later.

---

Mea culpa on the post title, but in my defense, I corrected with another post 4 days later...it would be a nice feature to let the original writer edit the thread title later.

You can. It's just conveluted. Once you post a new message, you edit that message and it gives you the topic to edit, also.

---

My movie https://vimeo.com/manage/videos/502242