![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
intel e6550 2.3 1333fsb 4mb cache @ Frys B&M for $179 + Tax
- Thread starter dslrjunky
- Start date
Temps aren't out of the question especially at the voltage. A low ambient temp (20C?) will give you really good temps.
well i just shutdown orthos, but it seems to be running @ 1.35 eventhough i went to the BIOS and put it @ 1.30..
also FWIW : I migrated from OPTY 185 > then e6600 w/ P5K DLX had a bunch of problems with it so i grab one of these GB P35-DQ6 and as soon as I sold my e6600 i bought a cheaper e6550...I didnt reformat the OS Drive I didnt reinstall windows all i did was load the P35 chipset Drivers for GB, i didn uninstall anything, i treated the mobo and proc as if they where just USB Drive PnP....
also FWIW : I migrated from OPTY 185 > then e6600 w/ P5K DLX had a bunch of problems with it so i grab one of these GB P35-DQ6 and as soon as I sold my e6600 i bought a cheaper e6550...I didnt reformat the OS Drive I didnt reinstall windows all i did was load the P35 chipset Drivers for GB, i didn uninstall anything, i treated the mobo and proc as if they where just USB Drive PnP....
CoreTemp only reports the VID which is the default voltage for your processor. The first page of CPUz should report your actual core voltage correctly.
Your temperatures are definitely believable. I took my E6400 and dropped the multi down to 7 and set my bios voltage to 1.30 volts like you did for a comparison.
My case is closed so I cranked up the Tuniq to full speed. Room temperature is 24C. Low voltage = cool CPU.
The old revision B2 wasn't such a bad processor.
Your temperatures are definitely believable. I took my E6400 and dropped the multi down to 7 and set my bios voltage to 1.30 volts like you did for a comparison.
My case is closed so I cranked up the Tuniq to full speed. Room temperature is 24C. Low voltage = cool CPU.
The old revision B2 wasn't such a bad processor.
I definitely like my Gigabyte 3D Aurora 570 case. It has two slow speed 120mm fans on the back and lots of space in there. I like silence so I turned off the front 120mm fan.
The best way to control cpu heat is to turn down the voltage as low as possible. I did the same thing with my x1950XT which made a big difference. A couple of 8800GTX cards in SLI is almost enough heat to do some baking in there while gaming away!
Hopefully iseedeadpeople will keep pushing his chip so we can make a more or less fair comparison between the new G0 stepping and last year's B2. I'm pretty sure the G0 is better but a lot of posts on XS, etc. have been with engineering sample (ES) processors. I want to see if the retail G0 processors are just as good.
The best way to control cpu heat is to turn down the voltage as low as possible. I did the same thing with my x1950XT which made a big difference. A couple of 8800GTX cards in SLI is almost enough heat to do some baking in there while gaming away!
Hopefully iseedeadpeople will keep pushing his chip so we can make a more or less fair comparison between the new G0 stepping and last year's B2. I'm pretty sure the G0 is better but a lot of posts on XS, etc. have been with engineering sample (ES) processors. I want to see if the retail G0 processors are just as good.
Nice results! I can't believe that ram rated for 400mhz is runing 540 mhz! You should be at the extreme end of the ram's limits. So it's probably going to be the first reason you'll get errors from that point if you don't raise the ram voltage or relax timings more. Again you should knock that Ram multiplier down to 2x to make make it run at 450mhz along with the FSB since that ram can obviously do DDR900 w/no issue. I'd would think from this point there may need to be a gradual cpu voltage increase. I've got to get my hands on one of these chips. Amazing so far. Great job guy!
Regarding Coretemp: I used to do some testing with some very accurate probes and on my system coretemp frequently was inaccurate at temps lower than 40c usually displaying 5-8c lower than actual. I noticed that above 40c actual it was usually dead on.
Regarding Coretemp: I used to do some testing with some very accurate probes and on my system coretemp frequently was inaccurate at temps lower than 40c usually displaying 5-8c lower than actual. I noticed that above 40c actual it was usually dead on.
Finally!!!! I've noticed the exact same thing happening but trying to convince others has been nearly impossible.
I've done plenty of testing as well and have compared CoreTemp to the output from the on chip digital thermal sensors by reading registers within the processor using CrystalCPUID. CoreTemp exactly tracks changes in the DTS which leads me to believe that core idle temperatures are not accurate because the digital thermal sensors are not accurate at low temperatures. This might be why Intel does not document the DTS to be used to report absolute core temperatures across the entire temperature range of a Core 2 Duo.
I found CoreTemp to be very accurate from about 45C to 85C as compared to an IR thermometer. Idle temperatures are meaningless if the sensor is not accurate at lower temps.
At the load temperatures you're running now, CoreTemp 0.95 and SpeedFan 4.32 are very accurate. Avoid bios temperatures.
Idle temperatures aren't important either. The lower the better as reported by one of these programs but comparing them to other users when there seems to be some error in the sensor that they are coming from is pointless.
Drop you memory multi to 2 to get your memroy running at 450 MHz and keep bumping up the FSB MHz. We all want to see what she'll do!
Idle temperatures aren't important either. The lower the better as reported by one of these programs but comparing them to other users when there seems to be some error in the sensor that they are coming from is pointless.
Drop you memory multi to 2 to get your memroy running at 450 MHz and keep bumping up the FSB MHz. We all want to see what she'll do!
I'll try and do that later, I'm doing this remotely actually I was gonna bring it to work but figured out working as an RN & Data Support i wouldnt have enough time to do all this, plus all the JCAHO crap thing would be thrown at my face when my CIO sees me in my office wiht flashing LED LIGHTS!!!...At the load temperatures you're running now, CoreTemp 0.95 and SpeedFan 4.32 are very accurate. Avoid bios temperatures.
Idle temperatures aren't important either. The lower the better as reported by one of these programs but comparing them to other users when there seems to be some error in the sensor that they are coming from is pointless.
Drop you memory multi to 2 to get your memroy running at 450 MHz and keep bumping up the FSB MHz. We all want to see what she'll do!
Like unclewebb, I feel the load temps would be accurate. Most bios's I've used on the contrary to him were pretty accurate except for one board (against my testing). Depends on the board. The idle coretemp readings should be taken with a grain of salt. You can always verify it's accuracy once you know the system well by lowering the cpu voltage to the minimum along with the cpu mhz. On a cold, unplugged computer, you would boot up, go into bios and check the temp against ambient. Unplug, wait another hour, startup, go into windows and immediate check coretemp. If it reports under ambient then you will know it's off at idle. You really don't need to worry about verifying load temps until you start to increase the cpu voltage +15% above stock voltage when at load. From what I've seen coretemp was always within 2c of my external tests with a good probe on the naked die. That was from 40c to the highest I brought the my opty chip to 65c.
So far it looks like you might surpass 3.5ghz and redeem in on a good 3dmark performance score especially once that video card is overclocked. I wonder if you might end up getting close to 15k on it.
So far it looks like you might surpass 3.5ghz and redeem in on a good 3dmark performance score especially once that video card is overclocked. I wonder if you might end up getting close to 15k on it.
Here's the reason I lost faith in bios temperature readings.
That 13C is significantly below ambient temperature which is impossible when air cooled. After I booted up SpeedFan reported that sensor as CPU = 12C.
Core temperatures can vary up or down by as much as 25C within one second as you go from no load to full load or from full load to no load so don't try to re-boot and look at your bios temps because it won't tell you anything.
CoreTemp or SpeedFan are very accurate for reporting core temperatures at full load so keep using either one of them. You have a pile of temperature head room, you're memory is very willing, and 7 X 500 MHz is quite likely and well within reach with only a minor voltage tweak or two.
That 13C is significantly below ambient temperature which is impossible when air cooled. After I booted up SpeedFan reported that sensor as CPU = 12C.
Core temperatures can vary up or down by as much as 25C within one second as you go from no load to full load or from full load to no load so don't try to re-boot and look at your bios temps because it won't tell you anything.
CoreTemp or SpeedFan are very accurate for reporting core temperatures at full load so keep using either one of them. You have a pile of temperature head room, you're memory is very willing, and 7 X 500 MHz is quite likely and well within reach with only a minor voltage tweak or two.
download "systool"
It should be able to identify most of you board. It has a nvidia tab and you can overclock both gpu and memory there on the fly. Put it at 610 gpu and 950 memory to start. The score should jump. If you can do over 625 on the core and 1050 on the memory then that card is nice. If systool recognizs your board well then you can adjust the fsb there too. But at the same time it will be increase that ram as well. Systool will ask if if can send updates when you install it and ask for a dozen or so sensor readers. Don't check those boxes or it may lock the computer up.
It should be able to identify most of you board. It has a nvidia tab and you can overclock both gpu and memory there on the fly. Put it at 610 gpu and 950 memory to start. The score should jump. If you can do over 625 on the core and 1050 on the memory then that card is nice. If systool recognizs your board well then you can adjust the fsb there too. But at the same time it will be increase that ram as well. Systool will ask if if can send updates when you install it and ask for a dozen or so sensor readers. Don't check those boxes or it may lock the computer up.
Well your CPU is really pushing the RAM because of its low multiplier. Right now you've gotten the CPU up to 3.5Ghz and the RAM has run at DDR1066 speeds.
In my opinion, you should be very happy your DDR2-800 overclocked to those speeds.
So I think your RAM might be holding you back. If you want to make sure, raise your memory divider and then continue raising the FSB.
In my opinion, you should be very happy your DDR2-800 overclocked to those speeds.
So I think your RAM might be holding you back. If you want to make sure, raise your memory divider and then continue raising the FSB.
it has been fun and all, but i gotta enjoy this machine now need to play some games and stop benching.......atleast ive learn some OC tips from you guys!!! thanks.... im happy i got this cpu...i was wondering how far the e6850 can go ....but who knows??? C2Quad is only 5 days away from the price cut!!
Time for an upgrade to SuperPI mod. It uses 3 digits after the decimal point.
Hopefully you can show us the CPU tab in CPUz when you are running Orthos at 3500 MHz and 1.30 volts. Do it for 4 hours and you can get your name in the [H]ard OC database and be famous.
Hopefully you can show us the CPU tab in CPUz when you are running Orthos at 3500 MHz and 1.30 volts. Do it for 4 hours and you can get your name in the [H]ard OC database and be famous.
well you urself us famous already,my 3.5 is nothing compare to your 3.6
I guess the problem I have is that I don't actually see your 3.5. You've posted some screen shots of some programs that are known to mis-report MHz. How about post one with CPUz showing your CPU at 3.5 GHz, preferably while running Orthos?
I don't have any trouble backing up my BS.
http://www.hardforum.com/showpost.php?p=1031178478&postcount=670
Edit: Me stupid for doubting issdeadpeople. His overclock is the real deal.
Gaming will be the real test. Do that at the overclock w/o overclocking the card yet to get a feel for stability. Notice how your cpu score jumped. Once you overclock that card and squeeze that last bit of mhz out of that cpu you'll hit that 15k. You should be getting close to instability after that point (however you still got vcore room, a plus). I would take anything over 3.7ghz as a Godsend. Excellent memory too! I bet that system screams on FPS games. You should try it out on Cellfactor Revolution and see if that game lags on that machine.
For comparison here is my system at it's peak (3ghz opty 170 / GTX @ 625/1050) watercooled cpu
sm2 score 4950
Hdr/sm2 core ~5020
Highest score 11,710
across the board your getting about 20% better performance on that chip than what I can get (what I've been interested in). That is with your card not even overclocked yet. that is the difference between getting a noticable 40fps minimum to 48fps which is much less noticable. nice.
ah ok i see what you mean, but like i said im not here to prove/claim or at best make a name for myself why is it i feel like your trying to say something?? anyway like i said as much as i enjoy this benchmarking overclocking wont serve the purpose of this rig...thanks..
CoW]8(0);1031275305 said:
Thanks CoW]8(0) for this suggestion. CPUz is one of the very few programs that displays the multiplier used correctly. If someone was pretending that their computer was running at 500 x 7 =3500 MHz and it was really only running at 500 x 6 = 3000 MHz, CPUz would clearly show that deception.
You're absolutely right. Your 3.5 is nothing because so far it hasn't been proven. I wasn't hoping to see a 5 hour Orthos run but 5 minutes with CPUz showing your actual MHz would have been nice.
Edit: Me really stupid for doubting issdeadpeople. His overclock is the real deal.
thats why i put in both cpuz and core temp in there, because i know that for some reason when the PC is in idle the clock goes down.... i posted the clocks that i did because i wanted to know how to OC not to claim i have an uber clocker, because i would love to be proved wrong when it comes to OC'ng as i dont know jack about overclocking... so i dont know why suddenly you turn into a jackass saying all this things.....
check post #68 and see what im saying, cpuz says 3.0 but on core temp it says 3.5(500x7) i run 3dmark06 after that it clocks back to 3.0 but when it runs super pi it shows 3.5 my ram freq. is 500
I apologize for being a dink. Everyone here is anxious to know exactly what these new processors are capable of and at what voltage.
If you have C1E or EIST ( Enhanced Intel SpeedStep Technology ) enabled in the bios then your multiplier will drop to 6 which lowers your MHz when your processor is idle. If you run Orthos then your cpu will be at full load and if you run CPUz beside that and take a screen shot then we can see the actual MHz your computer is running at as well as the voltage it requires to be stable at that MHz.
Your SuperPI and 3DMark CPU score seem to show you really are running at over 3500 MHz but I just needed to see some proof before running out and buying one of these new super chips. 3570 MHz at default voltage is incredible. My old revision B2 is no match for that.
If you have C1E or EIST ( Enhanced Intel SpeedStep Technology ) enabled in the bios then your multiplier will drop to 6 which lowers your MHz when your processor is idle. If you run Orthos then your cpu will be at full load and if you run CPUz beside that and take a screen shot then we can see the actual MHz your computer is running at as well as the voltage it requires to be stable at that MHz.
Your SuperPI and 3DMark CPU score seem to show you really are running at over 3500 MHz but I just needed to see some proof before running out and buying one of these new super chips. 3570 MHz at default voltage is incredible. My old revision B2 is no match for that.
np! was just surprised to your reaction on my oc results....i for myself is still in doubt about how good oc'er this is as i have no experience whatsoever and suddenly i reached 3.5ghz so i know i might be doing something odd or the board itself is the oc'ng beyond the fsb and multiplier i adjusted.... regarding the speedstep for some reason i cant find that option, i know how to find it on an asus BIOS but not on a gigabyte... this is my first time dealing with Intel, Gigabyte and DDR2...as I have been a loyal 754&939 fanboi....
was just surprised to your reaction on my oc results....i for myself is still in doubt about how good oc'er this is as i have no experience whatsoever and suddenly i reached 3.5ghz so i know i might be doing something odd or the board itself is the oc'ng beyond the fsb and multiplier i adjusted.... regarding the speedstep for some reason i cant find that option, i know how to find it on an asus BIOS but not on a gigabyte... this is my first time dealing with Intel, Gigabyte and DDR2...as I have been a loyal 754&939 fanboi....