Sign in with
Sign up | Sign in
Your question

I think I have a problem

Last response: in Systems
Share
January 18, 2012 1:47:27 PM

But I'm not sure what it is.

I built a new PC.

Case: Fractal design R3
Motherboard: Intel DZ68BC
CPU: Intel i5-2500K
Cooler: RTS2011LC liquid cooler
Memory: 16GB Gskill DDR3 -1333 7-7-7-21
Hard Drive: Intel RST ( 20GB Intel 311 SLC SSD and a 500GB Hitachi 7,200 hard drive)

Now I'm trying to check it out.

I ran the Mersenne Prime number search program in torture test mode. All 4 cores stressed to 100% if I look at task manager. But CPU temp and fan speed did not increase at all. SO I'm wondering if something is throttling my setup?

Here's my desktop: http://i42.tinypic.com/16gbblz.jpg

I need to find a benchmark that when I run it I can compare it with other i5-2500K CPU benchmarks results. ANyone know what I can run?

Anyone know why my temps aren't going up even when I stress things?

More about : problem

a b å Intel
January 18, 2012 1:50:28 PM

Hello Karsten75;

What were the CPU temps before and during the Mersenne Prime benchmark run?
m
0
l
January 18, 2012 1:52:26 PM

Doesn't sound like there is a problem, it is probably just the CPU cooler doing what it is supposed to do.
m
0
l
Related resources
January 18, 2012 2:02:21 PM

It's pretty cold in my house today - I'm not running A/C to heat it up. :) 

During idle, the CPU temps drop to 23C. When I run the Windows Performance index and the graphics (I use the integrated graphics) the temp climbs to 34C briefly and then drops to 30 for the rest of the test.

Thing is, I wasn't expecting to be this successful in building a super-quiet PC -the fans on this thing barely goes over 800RPM.
m
0
l
January 18, 2012 2:10:15 PM

I think I'm gonna have to try the cooler youre using :-)
m
0
l
a b å Intel
January 18, 2012 2:21:18 PM

Temps sound great. For idle of 23 C, room temp is around 68 F I'm quessing.

Try running prime 95 to stress CPU (Also checks stability of CPU/Memory)
Provide individual core temps - My quess is they will also be great.

probably the best gauge of performance would be to compare the time it takes to load windows 7 (time from "Start loading -> end of that spinning circle). Just for comparision My I5-2500k w/Samsung 830 is 12 Sec and for I5-2410m laptop w/m4 it is 17 sec. This is a better gauge than Synthetic benchmarks.
m
0
l
January 18, 2012 2:30:08 PM

RetiredChief said:
Temps sound great. For idle of 23 C, room temp is around 68 F I'm quessing.

Try running prime 95 to stress CPU (Also checks stability of CPU/Memory)
Provide individual core temps - My quess is they will also be great.

probably the best gauge of performance would be to compare the time it takes to load windows 7 (time from "Start loading -> end of that spinning circle). Just for comparision My I5-2500k w/Samsung 830 is 12 Sec and for I5-2410m laptop w/m4 it is 17 sec. This is a better gauge than Synthetic benchmarks.


Mersenne Prime (which is Prime 95) (which I ran) couldn't get the temps up. It seems that if I stress the HD 3000 integrated graphics, temps go up, otherwise not much.

Edit: yes, room temps are probably 68 or a little less, that room is the last to get heat and faces north.
m
0
l
January 18, 2012 3:18:31 PM

Just some update. I ran Passmark and that gave crappy results, some of which can be attributed to the lack of a video card, I think.

Now I"m running the Intel Extreme Tuning utility to see if and how much I can overclock the CPU. That got temps up to 46C. :) 
m
0
l
a b å Intel
January 18, 2012 3:18:52 PM

My goof, very few refer to prime 95 as Mersenne Prime, temporaty amnesia.

Is that 34 C cpu temp ore the average of the individual cores. Myself I normally ignore the CPU temp and just monitor the cores.

Added: Is see you are using ram very simular to Mine DDR3-1600 CL 7 @ 1.60 Volts.(also on I5-2500k)
Might take a read at this on using ram above 1.575 V:
http://www.tomshardware.com/forum/forum2.php?config=tom...
Read last update By Intelguy.
m
0
l
January 18, 2012 3:36:39 PM

Boot times are bad tests of how good a computer is.

It is too hard to control various hardware and settings factors.

All that a boot time benchmark is good for is showing how good you are at optimizing boot time and/or how much money you are willing to spend to get a good score.

You could have 2 different systems that perform the exact same inside the OS and one of them takes 15 more minutes to boot up than the other.
m
0
l
January 18, 2012 3:38:17 PM

RetiredChief said:
My goof, very few refer to prime 95 as Mersenne Prime, temporaty amnesia.

Is that 34 C cpu temp ore the average of the individual cores. Myself I normally ignore the CPU temp and just monitor the cores.

Added: Is see you are using ram very simular to Mine DDR3-1600 CL 7 @ 1.60 Volts.(also on I5-2500k)
Might take a read at this on using ram above 1.575 V:
http://www.tomshardware.com/forum/forum2.php?config=tom...
Read last update By Intelguy.


I believe I was very careful with my memory. This is the exact memory I bought. It is at 1.5V and at 1333Hz. Which is all within spec, I hope. I agonized a lot over the memory selection. G-Skill is an unknown for me.
m
0
l
January 18, 2012 3:44:38 PM

Raiddinn said:
Boot times are bad tests of how good a computer is.

It is too hard to control various hardware and settings factors.

All that a boot time benchmark is good for is showing how good you are at optimizing boot time and/or how much money you are willing to spend to get a good score.

You could have 2 different systems that perform the exact same inside the OS and one of them takes 15 more minutes to boot up than the other.


Who said anything about a boot time benchmark?
m
0
l
a b å Intel
January 18, 2012 3:59:05 PM

When I posted your link, it took me to gskill and the 1600 speed, but you clearly indicated 1333. Just looked at new egg and see it as relatively new. May try lowering my voltage Have 16 gigs so a little expensive to replace all of it.. But then my 1600 Cl 7 is just 0.025 (1.6%) above that limit. Very little diff between 1333 and 1600 for SB.
m
0
l
January 18, 2012 4:14:37 PM

Karsten75 said:
Who said anything about a boot time benchmark?


RetiredChief said:
probably the best gauge of performance would be to compare the time it takes to load windows 7


That guy.
m
0
l
January 18, 2012 5:23:27 PM

Raiddinn said:
That guy.


Sorry, I completely missed that. :(  My bad. And yes, FWIW, I agree that checking boot times is not a good gauge of performance.
m
0
l
a b å Intel
January 18, 2012 7:08:00 PM

Both are correct on Boot time dependent on system configuration, However
pray tell me why you think "Synthetic" Benchmarks are good - other than you like to see high sequencial performance, which is the LEAST important matrix for a OS + program drive or benchmark that uses highly compressed data. The only benchmark I look at anymore is PCMark vantage which is closer to real life than the others. At least AS SSD uses compressted data.

If the timing of loading windows is done after win 7 installation, driver and OS updates and Prior to program loads system programs would not be a factor. The primary facytors would be the SSD and indivdual choice of Hardware such as HDD controller chipset.
m
0
l
January 18, 2012 7:38:39 PM

That's a good question. :) 

But you have to admit that booting up a computer hardly strains it. It is mostly waiting for drivers and programs to load and initialize.

In my case what I want to do is to ensure that I can in fact drive the 4 cores I have to their maximum possible utilization (however artificial that is). I want to do this to make sure the cooling I installed is fully functional and keeps the processor cool enough. I want to make sure the processor does not throttle back due to temperature or other issues.

The only way I know to do this is to run a benchmark and compare it with other benchmarks.
m
0
l
a b å Intel
January 18, 2012 7:45:07 PM

Sorry, Must have got off topic, as I was addressing the SSD performance

Prime 95 should stress your all four cores of your CPU to verify your cooling system.
Normal systyem usage, ie encoding video file, or gaming will not heat your CPU up as much as prime 95 - which seems to be the "standard" here to verify cpu temps/cooling eff.
m
0
l
January 18, 2012 7:46:48 PM

Synthetic benchmarks, or benchmarks that are designed to stress one component and do nothing with another, are not something that translates closely to the experience using regular programs.

The benchmarks I prefer are the ones that mirror things people would do in RL, like if a company plays a game with a series of video cards and records the FPS in each case and then compares them.

I don't really see how the statistic of how long it takes to load Windows before installation of drivers and updates and programs is any sort of useful statistic even if you could measure it. If anything that is more of a synthetic benchmark because it in no way reflects the performance you will have under regular conditions.

That is unless you want your PC to remain with old drivers, no updates, and no programs loaded.

m
0
l
a b å Intel
January 18, 2012 7:59:15 PM

^ If you caught it, I stated that the ONLY bench marks I pay attention to are PCMark vantage, as tha is the closest to real life.

But the Only thing that is Important is How the (in this case) the SSD performs on the Specific system with what is installed on that system and the Hardware that is used. The inital run is only for comparision sake an you can compare that to other systems, as you modify system/update drivers ect then you are concerned about the effect on that individual compute - and that is done by user, in real life working on the system - not comparing to other systems. If gamer he/she can run demos, But here again it is what the user sees in real life run the game/ productivity software.

Benchmarks for SSD have become more about marketing than "stressing" the SSD
m
0
l
January 18, 2012 8:27:28 PM

I don't even really know why we are trying to stress a SSD in the first place.

That is probably my problem.
m
0
l
January 27, 2012 3:17:43 AM

Karsten75 said:
But I'm not sure what it is.

I built a new PC.

Case: Fractal design R3
Motherboard: Intel DZ68BC
CPU: Intel i5-2500K
Cooler: RTS2011LC liquid cooler
Memory: 16GB Gskill DDR3 -1333 7-7-7-21
Hard Drive: Intel RST ( 20GB Intel 311 SLC SSD and a 500GB Hitachi 7,200 hard drive)


What's your PSU???
m
0
l
!