- Status
dirtdiver
Distinguished
- Oct 26, 2009
- 284
- 0
- 18,790
IIRC a "good" E8400 will get moved into the E8500 or E8600 bin during manufacturing since they are based on the same architecture. My E8400 at 4.05ghz 1.3v would probably also run 4.4ghz...I guess I will have to give it a shot--but since I run 24/7 I like to keep it idling under 40c which has been my main reason for staying at 4.05ghz.
chkrdflg
Distinguished
- Dec 11, 2009
- 1
- 0
- 18,510
I am looking to build a new system. I was going to use a 4890 along with the Phenom II X2 550. I play driving sims GTR2, Richard Burns Rally, NASCAR 2003 etc. These games are actually older now. It looks like two of the games that benefit from more processor are the the driving games. Do you think it would benefit to jump up to the Phenom II X3 720? How do I know if the games that I play are more CPU or GPU dependent?
Greg Ritchie
Distinguished
- Apr 7, 2009
- 16
- 0
- 18,510
Hope I haven't missed the last couple parts of this (III and IV?). I really wanted to see the OC results.
My E8400 is CO stepping and a bit temperamental about voltage. It'll run 3.6GHz all day with 1.4V (on Win 7 64 bit, and 4.0GHz on XP Pro stably), but requires 1.425v to be stable at 4.0GHz (on Win 7 for whatever reason), and has reached 4.45GHz at 1.5v (all on air, just a cheap TT BT with TT 2500RPM fan). I've tested several hundred of these CPUs on systems I've built and get roughly the same results on all of them, regardless of stepping. So I'll be a bit disappointed if TG can't use at least 1.45v to overclock this dog. Why limit the voltage to less than a well proven safe voltage?
My E8400 is CO stepping and a bit temperamental about voltage. It'll run 3.6GHz all day with 1.4V (on Win 7 64 bit, and 4.0GHz on XP Pro stably), but requires 1.425v to be stable at 4.0GHz (on Win 7 for whatever reason), and has reached 4.45GHz at 1.5v (all on air, just a cheap TT BT with TT 2500RPM fan). I've tested several hundred of these CPUs on systems I've built and get roughly the same results on all of them, regardless of stepping. So I'll be a bit disappointed if TG can't use at least 1.45v to overclock this dog. Why limit the voltage to less than a well proven safe voltage?
Um.. I'm a pretty successful system builder. I've built almost a thousand systems with the E8400, and most react more like mine than yours with voltage. The E8400 is a great overclocker, but its also well known to be a bit voltage hungry. Most will require 1.40-1.45v to get above 4GHz, and usually 1.425-1.45v to be (orthos) stable with a stock cooler or cheap aftermarket cooler. But we don't install high end coolers like yours either, so that might make a little difference.
Honestly, most CPUs that perform similarly to yours would have been binned up to E8500. ;-)
Mine has been running fine at 4.0GHz on air for over a year, and is about ready to be retired (switching over to i7 860 system). But for whatever reason, it requires slightly more voltage on Win 7 than on XP to be stable, and that's what I've seen commonly on the systems we've built.
After building, we overclock and benchmark every system to be sure its stable and record the results, before putting it back to factory settings (in case our customers try to OC, so we know what they can and can't do). In fact, we've only used 2 motherboards for the most part, Gigabyte's EP45-DS3P and EP45-UD3R. I tell anyone wanted to OC to 4GHz to simply raise FSB to 445x9, raise MCH .2v, and set CPU voltage to 1.425v. If its not stable, increase CPU voltage to 1.45v. 19 out of 20 times that is all that is needed. The few exceptions need a better cooler.
Here's mine, and something for you to shoot for.
http://www.ultramaxcc.com.au/images/445GHz.jpg
Honestly, most CPUs that perform similarly to yours would have been binned up to E8500. ;-)
Mine has been running fine at 4.0GHz on air for over a year, and is about ready to be retired (switching over to i7 860 system). But for whatever reason, it requires slightly more voltage on Win 7 than on XP to be stable, and that's what I've seen commonly on the systems we've built.
After building, we overclock and benchmark every system to be sure its stable and record the results, before putting it back to factory settings (in case our customers try to OC, so we know what they can and can't do). In fact, we've only used 2 motherboards for the most part, Gigabyte's EP45-DS3P and EP45-UD3R. I tell anyone wanted to OC to 4GHz to simply raise FSB to 445x9, raise MCH .2v, and set CPU voltage to 1.425v. If its not stable, increase CPU voltage to 1.45v. 19 out of 20 times that is all that is needed. The few exceptions need a better cooler.
Here's mine, and something for you to shoot for.
http://www.ultramaxcc.com.au/images/445GHz.jpg
pauldh
Illustrious
- Jan 1, 2004
- 11,814
- 0
- 40,810
No, you haven't missed part 3 & 4. The last of the new hardware just trickled in, so we can confirm in part 3 you'll see Core i5, a few Radeon HD 5000's, and a switch to Win 7.
1.45V has been Tom's normal limit for a Wolfdale. I've pushed 1.5V myself, but not in systems used long enough to determine a life span of the CPU under such conditions. We'd rather play it safe for this series and not push beyond what readers could hit (or would run) at home with a similar config. Based on past reader feedback, 4.5 GHz @ 1.5V would be a bit meaningless for most readers.
One sample in the lab lost a significant amount of OC headroom after running 1.45V on a daily basis, so there's reason to recommend a bit less voltage. http://www.tomshardware.com/reviews/overclock-phenom-pentium,2366-10.html
Likely 1.4V will be the limit used, but I won't decide for sure until seeing the chip's scaling. Expectations are at least 4.0GHz, be it 1.35 or 1.4V
Thanx Pauldh and very cool, can't wait to see the results. I think the i5 750 will turn some heads with results too. I have yet to read anything about the new i5's and i3's.
That link points to a review on an E5200 though, which mentions an "office" system, and refers to Core 2 architecture in general for 1.45v being considered "safe", but doesn't state which specific CPU it was. So not sure if that's apples and oranges. I would imagine that there would be differences from 1 model to another when multipliers and cache and even bus rates vary. But even it if was a Wolfdale, that would seem to be a fluke from my experiences with them.
Just to be clear, I wasn't suggesting TG use 1.5v, as that was directed at Dirtdriver. And for the record, that particular 1.455GHz was achieved after running that E8400 @ 1.45v and 4.005GHz for over a year, and its run fine at 1.45v @ 4.0GHz since (a few months ago).
I think 1.45v is realistic for the tests as we've run quite a few of these CPUs at that 24/7 for quite some time now (14-18 months so far, and obviously rebooting every week or 2). I've seen huge differences in stability at 1.425v and 1.45v on various Wolfdales at 4GHz and over, and occasionally one will perform fine at 4.1GHz on 1.4v. I guess that's to be expected as no 2 CPUs are exactly alike. Obviously I don't know if they'll last for 3 years or not either, but so far so good.
Here's to hoping you guys decide on at least 1.425v or 1.45v, or get a very good sample to work with that'll perform well on less.
That link points to a review on an E5200 though, which mentions an "office" system, and refers to Core 2 architecture in general for 1.45v being considered "safe", but doesn't state which specific CPU it was. So not sure if that's apples and oranges. I would imagine that there would be differences from 1 model to another when multipliers and cache and even bus rates vary. But even it if was a Wolfdale, that would seem to be a fluke from my experiences with them.
Just to be clear, I wasn't suggesting TG use 1.5v, as that was directed at Dirtdriver. And for the record, that particular 1.455GHz was achieved after running that E8400 @ 1.45v and 4.005GHz for over a year, and its run fine at 1.45v @ 4.0GHz since (a few months ago).
I think 1.45v is realistic for the tests as we've run quite a few of these CPUs at that 24/7 for quite some time now (14-18 months so far, and obviously rebooting every week or 2). I've seen huge differences in stability at 1.425v and 1.45v on various Wolfdales at 4GHz and over, and occasionally one will perform fine at 4.1GHz on 1.4v. I guess that's to be expected as no 2 CPUs are exactly alike. Obviously I don't know if they'll last for 3 years or not either, but so far so good.
Here's to hoping you guys decide on at least 1.425v or 1.45v, or get a very good sample to work with that'll perform well on less.
I have to ammend the previous post:
All of our previously mentioned E8400s were bought - and the systems built - in Jan 2008 and March 2008. So the longest any of them has run is just under a year now. Sure seems longer.
The overclock that I posted was done on November 7th 2008, so that system has only run for about 2 months since then at 4GHz @ 1.45v, and ran that way for 10 months previous to that.
All of our previously mentioned E8400s were bought - and the systems built - in Jan 2008 and March 2008. So the longest any of them has run is just under a year now. Sure seems longer.
The overclock that I posted was done on November 7th 2008, so that system has only run for about 2 months since then at 4GHz @ 1.45v, and ran that way for 10 months previous to that.
dirtdiver
Distinguished
- Oct 26, 2009
- 284
- 0
- 18,790
dark41 :
Um.. I'm a pretty successful system builder. I've built almost a thousand systems with the E8400, and most react more like mine than yours with voltage. The E8400 is a great overclocker, but its also well known to be a bit voltage hungry. Most will require 1.40-1.45v to get above 4GHz, and usually 1.425-1.45v to be (orthos) stable with a stock cooler or cheap aftermarket cooler. But we don't install high end coolers like yours either, so that might make a little difference.
Honestly, most CPUs that perform similarly to yours would have been binned up to E8500. ;-)
Mine has been running fine at 4.0GHz on air for over a year, and is about ready to be retired (switching over to i7 860 system). But for whatever reason, it requires slightly more voltage on Win 7 than on XP to be stable, and that's what I've seen commonly on the systems we've built.
After building, we overclock and benchmark every system to be sure its stable and record the results, before putting it back to factory settings (in case our customers try to OC, so we know what they can and can't do). In fact, we've only used 2 motherboards for the most part, Gigabyte's EP45-DS3P and EP45-UD3R. I tell anyone wanted to OC to 4GHz to simply raise FSB to 445x9, raise MCH .2v, and set CPU voltage to 1.425v. If its not stable, increase CPU voltage to 1.45v. 19 out of 20 times that is all that is needed. The few exceptions need a better cooler.
Here's mine, and something for you to shoot for.
http://www.ultramaxcc.com.au/images/445GHz.jpg
Honestly, most CPUs that perform similarly to yours would have been binned up to E8500. ;-)
Mine has been running fine at 4.0GHz on air for over a year, and is about ready to be retired (switching over to i7 860 system). But for whatever reason, it requires slightly more voltage on Win 7 than on XP to be stable, and that's what I've seen commonly on the systems we've built.
After building, we overclock and benchmark every system to be sure its stable and record the results, before putting it back to factory settings (in case our customers try to OC, so we know what they can and can't do). In fact, we've only used 2 motherboards for the most part, Gigabyte's EP45-DS3P and EP45-UD3R. I tell anyone wanted to OC to 4GHz to simply raise FSB to 445x9, raise MCH .2v, and set CPU voltage to 1.425v. If its not stable, increase CPU voltage to 1.45v. 19 out of 20 times that is all that is needed. The few exceptions need a better cooler.
Here's mine, and something for you to shoot for.
http://www.ultramaxcc.com.au/images/445GHz.jpg
I believe the large aftermarket cooler is likely the reason for the stable OC, and I believe I just got very lucky with the chip I bought
yah 445x9 is correct
I have since sold that system and bought an i7 860 so I guess I am starting over from scratch on the OC...interested to see how it does! Even more interested in parts III and IV of this article to be honest..
I'll be moving over soon too, but not sure yet whether i5 or i7, possibly even i3 but not likely. I've been playing with a few i5 750s lately (customer builds) and must say they're pretty impressive compared to my OC'ed E8400. But the BIOS is a totally new learning curve. The 860 is over $100 more here, so not sure if the bang for the buck is there, but hopefully this review will help shed some light on that as well.
dirtdiver
Distinguished
- Oct 26, 2009
- 284
- 0
- 18,790
@ Dark41,
I read the review on the new i5's on guru3d.com and I gotta say, they are impressive. For a dual core processor they run very cool, and are hyperthreaded so they can run those 2 extra logical cores. Very cool idea, and they overclock very nicely.
I'm in the same boat with learning the new BIOS for my 860, have only overclocked to (just a hair over) 3ghz, lol. So, essentially nothing. I just haven't had time to sit down and do the research on it yet.
Price/performance the i5 wins (unless you need hyperthreading). They OC very nicely from everything I've seen, and the next part of this articly should really show that to everone
I read the review on the new i5's on guru3d.com and I gotta say, they are impressive. For a dual core processor they run very cool, and are hyperthreaded so they can run those 2 extra logical cores. Very cool idea, and they overclock very nicely.
I'm in the same boat with learning the new BIOS for my 860, have only overclocked to (just a hair over) 3ghz, lol. So, essentially nothing. I just haven't had time to sit down and do the research on it yet.
Price/performance the i5 wins (unless you need hyperthreading). They OC very nicely from everything I've seen, and the next part of this articly should really show that to everone
Personally, I think over-clocking should be left out of the conversations and the article itself. I would be more interested in factual information based upon the hardware items purchased off the shelf. That's what I imagine that most users would want to know.
The resultant benefits/hazards of over-clocking should be left to the indiviual user's discretion AFTER they put together their system and plug it in at home.
Although the articles are extremely well-written and offer immense value (including the use of great, understandable graphics!), I began reading with the expectation of (hopefully) finding out how to spend the least amount of money to buy the best possible individual components.
The great article helped me, but after I spent much time reading almost every response (here), I'm now more confused than ever. There are too many "exceptions" and opinions on "alternatve hardware choices!" Plus, as I said before, the conversations on over-clocking confuse everything even more.
Everybody has the right to their own opinion... even me! Heh heh. But, I'm more interested in finding a proven, definitive list of items to buy.
Compare price/performance in the article, and then finally (after, or in, the conclusion)... present several, complete "Shopping lists" each of which itemizes the 'best combination of parts by price-range)! That's what I could use, especially if it had part and model number and estimated current price per item listed.
I'm too old to actually keep up with the myriad possible combinations in order to conclude what I should buy. Heh heh. Keeping it very simple would work best for me.
Footnote: Why are there no "Preview" or "Edit" options for these posts?
The resultant benefits/hazards of over-clocking should be left to the indiviual user's discretion AFTER they put together their system and plug it in at home.
Although the articles are extremely well-written and offer immense value (including the use of great, understandable graphics!), I began reading with the expectation of (hopefully) finding out how to spend the least amount of money to buy the best possible individual components.
The great article helped me, but after I spent much time reading almost every response (here), I'm now more confused than ever. There are too many "exceptions" and opinions on "alternatve hardware choices!" Plus, as I said before, the conversations on over-clocking confuse everything even more.
Everybody has the right to their own opinion... even me! Heh heh. But, I'm more interested in finding a proven, definitive list of items to buy.
Compare price/performance in the article, and then finally (after, or in, the conclusion)... present several, complete "Shopping lists" each of which itemizes the 'best combination of parts by price-range)! That's what I could use, especially if it had part and model number and estimated current price per item listed.
I'm too old to actually keep up with the myriad possible combinations in order to conclude what I should buy. Heh heh. Keeping it very simple would work best for me.
Footnote: Why are there no "Preview" or "Edit" options for these posts?
Kreelor,
These articles do a very good job of showing how these components perform at both stock and overclocked speeds. If you don't want to know about overclocking results, just skip past those parts/posts. You could also just use the charts TG provides to pick components that won't be overclocked. There's some very good resources here for those who want to run at stock settings.
But honestly, you're on the wrong site to suggest eliminating overclocking results (would start a riot if they stopped even, and probably drop their readship to virtually nothing ((pun intended)). This site is read mostly by enthusiasts who care very much about overclocking results (myself included). Heck, I'm 50 years young, and wish I was 15 years older so I could spend more time reading about these things and less time actually building them. Retirement will be greatly appreciated and used to do what I really enjoy, which is learning about the latest technology, and riding my BMW around the world.
These articles do a very good job of showing how these components perform at both stock and overclocked speeds. If you don't want to know about overclocking results, just skip past those parts/posts. You could also just use the charts TG provides to pick components that won't be overclocked. There's some very good resources here for those who want to run at stock settings.
But honestly, you're on the wrong site to suggest eliminating overclocking results (would start a riot if they stopped even, and probably drop their readship to virtually nothing ((pun intended)). This site is read mostly by enthusiasts who care very much about overclocking results (myself included). Heck, I'm 50 years young, and wish I was 15 years older so I could spend more time reading about these things and less time actually building them. Retirement will be greatly appreciated and used to do what I really enjoy, which is learning about the latest technology, and riding my BMW around the world.
Thanks for your reply. BTW, I AM that +15 years older that you mentioned. I'm 65.
I did not intend to imply that discussions of over-clocking should not be on THG. I was referring to the hardware comparisons. They were listed at the beginning of the article, if I remember correctly?
My view is that if the article compares specific hardware items (without mentioning over-clocking), the resultant article would still serve its purpose - to compare items!
That would make it a "fair" comparison, because not everything over-clocks the exact same way.
Here's what Paul wrote (about 8 or 9 paragraphs from the beginning of his article):
It's no big deal! I'm not really complaining. It's just that talking about over-clocking at this stage of the game is "way ahead of the game." THAT COMES AFTER this phase of the article(s).
I did not intend to imply that discussions of over-clocking should not be on THG. I was referring to the hardware comparisons. They were listed at the beginning of the article, if I remember correctly?
My view is that if the article compares specific hardware items (without mentioning over-clocking), the resultant article would still serve its purpose - to compare items!
That would make it a "fair" comparison, because not everything over-clocks the exact same way.
Here's what Paul wrote (about 8 or 9 paragraphs from the beginning of his article):
It's no big deal! I'm not really complaining. It's just that talking about over-clocking at this stage of the game is "way ahead of the game." THAT COMES AFTER this phase of the article(s).
- Status
TRENDING THREADS
-
News TSMC founder says Intel should focus on AI, not advanced process technologies- Started by Admin
- Replies: 30
-
-
-
-
Latest posts
-
Question Sensor problem with Logitech G Pro X Superlight 2 mouse ?- Latest: itzNighthawKx
-
-
-
Facebook
Twitter
Instagram