i actually looked at that one, the 600RPM on the low side put me off though, as weird as it sounds since i cant adjust the fan curve i was looking for a high low rpm.. i need to crack open this case and break out a ruler. ill end up doing that tomorrow, right now my allergies have me amazingly tired and worn out, definitely dont want to deal with dust bunnies right now.
aside: ive had amazingly bad luck with fans over the years, happen to know about how long the fan on the zalman might last? ive had fans die after 6 months or less (i was beginning to think i had some sort of curse on me or something lol)
and yeah, i spent the better part of the morning researching TIM's, to many mixed results, even the stuff that comes with it has mixed reviews. some folks saying it had "chunks" in the mix.. think ill just stick to what i know and trust, will be adding a new tube of AS5 to the shopping cart..
right now i have NNW set for Einstein. left coretemp on since yesterday it peaked at 171*f and is currently at 155*f..
I recently had a Zalman fan let go after over 5 years of near constant running. It is fixable, but not by me, a piece broke that holds the fan blades themselves into the circle of fins. If it hadn't broken it would still be running! I heard it tapping the blades when it was running when it was quiet one day, otherwise it could still be running, though broken.
I agree about the paste! Too much is NOT a good thing, I put a small blob, about the size of a pea, then spread it around before putting the cpu fan on it and locking it down. I have done it before without spreading and it seemed to run hotter, spreading works better for me.
1)
Did you get a chance to run the windows disk check?
( just something to rule out...)
2) heat
With your new fan settings, & thermal grease efforts,
what kind of temps have your CPU been running?
[ start gossip]
I run BOINC on Vista and on Linux. I did have a crash last year
when I did not clean out the dust on the heat sinks.
( I think it was good idea to change thermal grease.)
I live in Florida (getting hot here) -- have been setting BOINC to run on 75% of the cores.
I have 8 core cpu - and that lowered the temp about 3C.
I noticed less temp fluctuation that way than setting to run all cpu 75% of the time.
(I used piriform's "speccy" to see CPU temps - and they do a good graph).
Now, I run BOINC at night - when I can set the AC down to 75F.
I'm running 125 to 134F with just 75% (with 1 of those feeding the GPU)
[ end gossip]
3) External events
Anything else going on? Power surges? Anything weird in the event logs?
4) windows?
I must admit - I don't know windows 8 ...but...
Is a there still a task manager that can list page faults??
Are there over 1,000 p/f per period on the einstein apps? (that too much.)
5) overnight run?
How long did Einstein run before it crashed?
Maybe it would be worthwhile to do an OVERNIGHT run of one of the 'just try - do not install'
versions of linux and then run Einstein..
Debian and Ubuntu have the "psensor" package that can monitor CPU temps.
Arrrgh, problem with trying linux - may not help solve warranty argument. :(
Does it matter which of the einstein apps that you have selected?
- Binary Radio Pulsar Search (Arecibo)
- Binary Radio Pulsar Search (Arecibo, GPU)
- Binary Radio Pulsar Search (Perseus Arm Survey)
- Gravitational Wave S6 Directed Search (CasA)
- Gamma-ray pulsar search #3
and did you allow the Beta test code?
-(Might want to play it safe and not allow the beta test.)
May I be seated in the shoe-eating section of the dining hall, please?
Paul,
I did not check page faults on my system, until after I posted the above.
Oops.
On my old, 4 processor Vista system, I'm getting 185,000+ page faults a seconds.
With 4 processors, it has 5,886MB of memory.
It was having this high rate on the einstein application (Gravitational wave S6 Directed Search (CasA) 1.05 (SSE2)) after running for 10+ hours.
I tried running only 1,2,3 or all 4 processors
Still high page fault rates on 1 of the 4 applications.
I suspended BOINC, cleaned out trash files (CCleaner) defragmented and rebooted.
Then, I got TWO task with high page fault rates.
I watched this with Task manager for 10 minutes, and then, all of a sudden, no more page faults.
So. I suppose that this is normal for the app - and is probably why I ran Einstein on a PC with LOTS more memory.
I should correct myself and say that these might be 'soft' page faults -
but about 50 of the processor's time was in kernel tasks.
Even with the page faults, no system crash. I saw no unusual events in either the system or application logs.
I will try running Einstein CPU tasks on my PC that has lots of memory and see if there is/are page fault errors and look for crashes...
hey jay, task manager in 8.1 doesnt have that per task, at least not that i saw? i do know of performance monitor though, which can give a lot of information about a system.
i havent had a chance to test it out yet, got done with the install saturday. wanted to take it slow.
as for diskcheck. its part of my monthly diskimage, cleanup, routine. i doubt 8.1 really needs it, but i got in to the habbit a long time ago.
seeing without seeing is something the blind learn to do, and seeing beyond vision can be a gift.
ok so, my original plan wouldnt work. using a rather unscientific method (yard stick across the case and a ruler measuring from base of current heat sink to yard stick) i determined i did not have enough room for the one i wanted. in a way im glad because i was about to make a boo boo, that one only has a 3 pin connector, i needed a 4 pin, so i went with the "ZALMAN CNPS9500 AT" instead, i all so figured i should replace the case (system) fan, i went with a "Delta FFB0912SH-F00".
odd thing about that, even though there both 3 pin fans (dell installed and the new one) one has a yellow (or green?) wire, and the other has a blue (or purple?) wire. do those mean different functions?, in either case it fits. though it did give me a minor heart attack (more later)
this has been an adventure, i removed the motherboard to change out the plate, only to discover that the intel backplate (the entire socket) was not covered under the instructions, surprisingly, im not sure if this is a dell thing or what. but this backplate looks like this EK-LGA115x
so, the mounting screws would not go "through" it, but the threads as far as i could tell was a match for the previous heat sink. and after screwing them in they even protruded about the same amount. so i took the chance.
oh, as for thermal compound. i picked up some arcticlean (this stuff smells like super oranges!). surprisingly the previous TIM came off like powder, literally like fine talcom.
ive done my tim application the same way for years, i put a grain of rice size of CMQ-2 (arctic silver ceramique, sorry got that mixed up with as5 in the previous post) then i spread it out in all directions, then i use a credit card to scrape off excess, basically only leaving the voids filled. i do the same for the heat sink. doing this i noticed the cpu has a low spot in the middle a bit bigger than a dime but smaller than a nickle. and that the heatsink while mirror smooth has a "dip" on one side.
after getting those installed, it was time for a boot up. so i crossed my fingers and booted up, it started up and that delta fan sounded like a turbine at max, then of course it slowly reved down. as the system booted in to windows i suddenly heard a SNAP, uh oh :-| i shut down as fast as i could and popped the side open, the heat sink was still attatched and tight.
so i started it up with the case open and loaded the bios to look at fan rpms and temps, wouldnt ya know it? aparantly dell doesnt want you seeing that so its not in the bios :-|
so i booted in to windows and watched it. both fans worked, until windows starts, by then the case fan has reved down so low, that it cuts off :-|, unless its supposed to do that? i can hear a very very high pitch sound coming from it in a 1 second pulse on/off cycle.
i loaded up speed fan and gave it a try again. sure enough i could now rev up the fan, AND down the fan. the fan was showing as 20%, so thumping it to 25% got it going, though it did show 250k rpm's for a bit (yes, two hundred fifty thousand, not a typo)
so on the plus side i think i can program my own fancurve using speedfan now, if i can get that to autorun ill be set.
loaded up coretemp, and the temps have been in the 80 - 90*f range at idle, ive left it idle for a bit just to "be sure", so far its peaked to 130*f, odd thing though, i have one core thats consistantly about 5*f hotter than the other 3, im wondering if thats because of that dip, or just natural variation.
ive only had one hiccup so far, seems i messed up hooking up the 20pin usb 3.0 header, theres no "key" for it so i had to open her up and fix that.
as for that snap? my theory here is that since this heatsink is 100% copper, and copper expands faster than aluminum, that it was just an expansion pop. at least, thats what i HOPE it was. because so far shes running fine.
ill put her under load either today or tomorrow when i can watch her like a hawk and get back to you.
oh, as for those faults, it was all ways s6 casa in the memory dumps, and as for page faults i have no idea, ill try to figure out how to set up a performance counter for faults when i get Einstein back to running, but this system has 12GB's of memory, so it "shouldnt" be paging active tasks?
p.s. this heatsink is HUGE :)
seeing without seeing is something the blind learn to do, and seeing beyond vision can be a gift.
well, im thinking i messed something up, just not sure what. leaving it "as is" and running p95 results in 181*f (82*c), cranking the zalman to max with p95 on results in 170*f (76*c), cranking the case fan above 50% results in temps going UP instead of down, i think its choking out the zalman if i go high (theres only about 2 inches distance from case fan to heatsink), leaving that at 25-40% gets me no change on the cpu.
at idle its 86 - 90*f (30-32c)
so, im not sure if i put on "too little" TIM, or if even though zalmans site says the HSF fits LGA1150, that it "doesnt", or that oddball backplates a problem. not sure where to go from here.
thinking i may take the board out and make sure those screws didnt slip. but man thats a real pain. other wise i may just be SOL.
seeing without seeing is something the blind learn to do, and seeing beyond vision can be a gift.
ok, not pleased with zalman at the moment, this is the specs for this heatsink, it clearly shows socket 1150, i-7, all speeds. the reply from their tech support though is.
Quote:
CNPS9500 is suitable for i3 or some i5s. Since it is entry level cooler.
We highly recommend using highr performance cooler such as 9900 MAX for such a high performance CPU like i7-4770.
9500 is a great cooler and it does support 1150 but your CPU is way too high performance CPU to be cooled with entry level cooler.
not happy, at all. especially since the 9900max wont fit, well, not unless i cut a hole in the case. and since their site stated it would indeed work with this cpu.
::sigh:: looks like im just SOL until i can get a new case.
seeing without seeing is something the blind learn to do, and seeing beyond vision can be a gift.
ok, not pleased with zalman at the moment, this is the specs for this heatsink, it clearly shows socket 1150, i-7, all speeds. the reply from their tech support though is.
Quote:
CNPS9500 is suitable for i3 or some i5s. Since it is entry level cooler.
We highly recommend using highr performance cooler such as 9900 MAX for such a high performance CPU like i7-4770.
9500 is a great cooler and it does support 1150 but your CPU is way too high performance CPU to be cooled with entry level cooler.
not happy, at all. especially since the 9900max wont fit, well, not unless i cut a hole in the case. and since their site stated it would indeed work with this cpu.
::sigh:: looks like im just SOL until i can get a new case.
Can you just leave the side off until then? Or if you are going to replace it anyway cut a hole in the side and put a screen over the hole. The hole could be a good thing for the next build as it is then usable as a case fan mount that blows onto the cpu directly.
RE: RE: I love those
)
I recently had a Zalman fan let go after over 5 years of near constant running. It is fixable, but not by me, a piece broke that holds the fan blades themselves into the circle of fins. If it hadn't broken it would still be running! I heard it tapping the blades when it was running when it was quiet one day, otherwise it could still be running, though broken.
I agree about the paste! Too much is NOT a good thing, I put a small blob, about the size of a pea, then spread it around before putting the cpu fan on it and locking it down. I have done it before without spreading and it seemed to run hotter, spreading works better for me.
Hi there, 5
)
Hi there,
5 things.....
1)
Did you get a chance to run the windows disk check?
( just something to rule out...)
2) heat
With your new fan settings, & thermal grease efforts,
what kind of temps have your CPU been running?
[ start gossip]
I run BOINC on Vista and on Linux. I did have a crash last year
when I did not clean out the dust on the heat sinks.
( I think it was good idea to change thermal grease.)
I live in Florida (getting hot here) -- have been setting BOINC to run on 75% of the cores.
I have 8 core cpu - and that lowered the temp about 3C.
I noticed less temp fluctuation that way than setting to run all cpu 75% of the time.
(I used piriform's "speccy" to see CPU temps - and they do a good graph).
Now, I run BOINC at night - when I can set the AC down to 75F.
I'm running 125 to 134F with just 75% (with 1 of those feeding the GPU)
[ end gossip]
3) External events
Anything else going on? Power surges? Anything weird in the event logs?
4) windows?
I must admit - I don't know windows 8 ...but...
Is a there still a task manager that can list page faults??
Are there over 1,000 p/f per period on the einstein apps? (that too much.)
5) overnight run?
How long did Einstein run before it crashed?
Maybe it would be worthwhile to do an OVERNIGHT run of one of the 'just try - do not install'
versions of linux and then run Einstein..
Debian and Ubuntu have the "psensor" package that can monitor CPU temps.
Arrrgh, problem with trying linux - may not help solve warranty argument. :(
I hope you win the warranty race.
Jay
Well, One more
)
Well, One more question...
Does it matter which of the einstein apps that you have selected?
- Binary Radio Pulsar Search (Arecibo)
- Binary Radio Pulsar Search (Arecibo, GPU)
- Binary Radio Pulsar Search (Perseus Arm Survey)
- Gravitational Wave S6 Directed Search (CasA)
- Gamma-ray pulsar search #3
and did you allow the Beta test code?
-(Might want to play it safe and not allow the beta test.)
Good luck!!
Jay
May I be seated in the
)
May I be seated in the shoe-eating section of the dining hall, please?
Paul,
I did not check page faults on my system, until after I posted the above.
Oops.
On my old, 4 processor Vista system, I'm getting 185,000+ page faults a seconds.
With 4 processors, it has 5,886MB of memory.
It was having this high rate on the einstein application (Gravitational wave S6 Directed Search (CasA) 1.05 (SSE2)) after running for 10+ hours.
I tried running only 1,2,3 or all 4 processors
Still high page fault rates on 1 of the 4 applications.
I suspended BOINC, cleaned out trash files (CCleaner) defragmented and rebooted.
Then, I got TWO task with high page fault rates.
I watched this with Task manager for 10 minutes, and then, all of a sudden, no more page faults.
So. I suppose that this is normal for the app - and is probably why I ran Einstein on a PC with LOTS more memory.
I should correct myself and say that these might be 'soft' page faults -
but about 50 of the processor's time was in kernel tasks.
Even with the page faults, no system crash. I saw no unusual events in either the system or application logs.
I will try running Einstein CPU tasks on my PC that has lots of memory and see if there is/are page fault errors and look for crashes...
Jay
hey jay, task manager in 8.1
)
hey jay, task manager in 8.1 doesnt have that per task, at least not that i saw? i do know of performance monitor though, which can give a lot of information about a system.
i havent had a chance to test it out yet, got done with the install saturday. wanted to take it slow.
as for diskcheck. its part of my monthly diskimage, cleanup, routine. i doubt 8.1 really needs it, but i got in to the habbit a long time ago.
seeing without seeing is something the blind learn to do, and seeing beyond vision can be a gift.
ok so, my original plan
)
ok so, my original plan wouldnt work. using a rather unscientific method (yard stick across the case and a ruler measuring from base of current heat sink to yard stick) i determined i did not have enough room for the one i wanted. in a way im glad because i was about to make a boo boo, that one only has a 3 pin connector, i needed a 4 pin, so i went with the "ZALMAN CNPS9500 AT" instead, i all so figured i should replace the case (system) fan, i went with a "Delta FFB0912SH-F00".
odd thing about that, even though there both 3 pin fans (dell installed and the new one) one has a yellow (or green?) wire, and the other has a blue (or purple?) wire. do those mean different functions?, in either case it fits. though it did give me a minor heart attack (more later)
this has been an adventure, i removed the motherboard to change out the plate, only to discover that the intel backplate (the entire socket) was not covered under the instructions, surprisingly, im not sure if this is a dell thing or what. but this backplate looks like this EK-LGA115x
so, the mounting screws would not go "through" it, but the threads as far as i could tell was a match for the previous heat sink. and after screwing them in they even protruded about the same amount. so i took the chance.
oh, as for thermal compound. i picked up some arcticlean (this stuff smells like super oranges!). surprisingly the previous TIM came off like powder, literally like fine talcom.
ive done my tim application the same way for years, i put a grain of rice size of CMQ-2 (arctic silver ceramique, sorry got that mixed up with as5 in the previous post) then i spread it out in all directions, then i use a credit card to scrape off excess, basically only leaving the voids filled. i do the same for the heat sink. doing this i noticed the cpu has a low spot in the middle a bit bigger than a dime but smaller than a nickle. and that the heatsink while mirror smooth has a "dip" on one side.
after getting those installed, it was time for a boot up. so i crossed my fingers and booted up, it started up and that delta fan sounded like a turbine at max, then of course it slowly reved down. as the system booted in to windows i suddenly heard a SNAP, uh oh :-| i shut down as fast as i could and popped the side open, the heat sink was still attatched and tight.
so i started it up with the case open and loaded the bios to look at fan rpms and temps, wouldnt ya know it? aparantly dell doesnt want you seeing that so its not in the bios :-|
so i booted in to windows and watched it. both fans worked, until windows starts, by then the case fan has reved down so low, that it cuts off :-|, unless its supposed to do that? i can hear a very very high pitch sound coming from it in a 1 second pulse on/off cycle.
i loaded up speed fan and gave it a try again. sure enough i could now rev up the fan, AND down the fan. the fan was showing as 20%, so thumping it to 25% got it going, though it did show 250k rpm's for a bit (yes, two hundred fifty thousand, not a typo)
so on the plus side i think i can program my own fancurve using speedfan now, if i can get that to autorun ill be set.
loaded up coretemp, and the temps have been in the 80 - 90*f range at idle, ive left it idle for a bit just to "be sure", so far its peaked to 130*f, odd thing though, i have one core thats consistantly about 5*f hotter than the other 3, im wondering if thats because of that dip, or just natural variation.
ive only had one hiccup so far, seems i messed up hooking up the 20pin usb 3.0 header, theres no "key" for it so i had to open her up and fix that.
as for that snap? my theory here is that since this heatsink is 100% copper, and copper expands faster than aluminum, that it was just an expansion pop. at least, thats what i HOPE it was. because so far shes running fine.
ill put her under load either today or tomorrow when i can watch her like a hawk and get back to you.
oh, as for those faults, it was all ways s6 casa in the memory dumps, and as for page faults i have no idea, ill try to figure out how to set up a performance counter for faults when i get Einstein back to running, but this system has 12GB's of memory, so it "shouldnt" be paging active tasks?
p.s. this heatsink is HUGE :)
seeing without seeing is something the blind learn to do, and seeing beyond vision can be a gift.
RE: ill put her under
)
It'll be fun to know if all of that work fixed something.
well, im thinking i messed
)
well, im thinking i messed something up, just not sure what. leaving it "as is" and running p95 results in 181*f (82*c), cranking the zalman to max with p95 on results in 170*f (76*c), cranking the case fan above 50% results in temps going UP instead of down, i think its choking out the zalman if i go high (theres only about 2 inches distance from case fan to heatsink), leaving that at 25-40% gets me no change on the cpu.
at idle its 86 - 90*f (30-32c)
so, im not sure if i put on "too little" TIM, or if even though zalmans site says the HSF fits LGA1150, that it "doesnt", or that oddball backplates a problem. not sure where to go from here.
thinking i may take the board out and make sure those screws didnt slip. but man thats a real pain. other wise i may just be SOL.
seeing without seeing is something the blind learn to do, and seeing beyond vision can be a gift.
ok, not pleased with zalman
)
ok, not pleased with zalman at the moment, this is the specs for this heatsink, it clearly shows socket 1150, i-7, all speeds. the reply from their tech support though is.
not happy, at all. especially since the 9900max wont fit, well, not unless i cut a hole in the case. and since their site stated it would indeed work with this cpu.
::sigh:: looks like im just SOL until i can get a new case.
seeing without seeing is something the blind learn to do, and seeing beyond vision can be a gift.
RE: ok, not pleased with
)
Can you just leave the side off until then? Or if you are going to replace it anyway cut a hole in the side and put a screen over the hole. The hole could be a good thing for the next build as it is then usable as a case fan mount that blows onto the cpu directly.