QX6700 vs E8400

[WittieM]

Well I have a chance to get a new qx6700 from a friend for $300 or I could get a E8400/Xeon equivalent once they become available again. Which would you choose and why?

Thanks,
Matt

Comp will mainly be used for gaming if you need to know. That and wouldn't mind getting into folding.

[Richard Dower]

I'd get the QX6700 and overclock it, even a modest 3.2GHz would be faster then the E8400.

[RunawayPrisoner]

Hmm... QX6700 if you have adequate cooling. E8400 if you are really going to game that hard, and you never touch Crysis. :)

[PhilDoc]

Hmm... QX6700 if you have adequate cooling. E8400 if you are really going to game that hard, and you never touch Crysis. :)

Really, I'll take that challange. Your X3350 vs my e8400 in Crysis.

[RunawayPrisoner]

Really, I'll take that challange. Your X3350 vs my e8400 in Crysis.

Then prepare yourself. ;) I have a 8800GTS G92 right here. I think we can use the Crysis Benchmark tool for the bench. :) (And it also saves me the trouble of having to put my E8400 back in for a bench) Clock your E8400 as high as you can and then use the tool to obtain a result similar to this one:

http://i235.photobucket.com/albums/ee60/cuongvietpham/PC/CrysisBenchDX10.jpg

And we'll see. :) Oh, and the res should be either 1024 x 768 or 1280 x 1024 so our GPUs won't limit the bench... much. :)

[WittieM]

Well it will be either watercooled inside a TJ-10 case or under a phase changer in a Lian Li V2000 case. Although the new CM Cosmos S does look interesting.

BTW I await the results from both tests. :toast:

[3lfk1ng]

RunawayPrisoner use your 8800GTS G92 like you said and not a 9600gt.

I too will await the results from both you two.

[jaredpace]

lol oc'd e8400 owns all quads in games - crysis too.

quads own in folding & x264 & such

[3lfk1ng]

I agree Jared. I just want to see some friendly competition :)

[Dami3n]

For gaming forget the quads. They are great for benchs, video, 3d,... but in games, you will have two cores sleeping while other two are working(and sometimes only one) :D

[PhilDoc]

Then prepare yourself. ;) I have a 8800GTS G92 right here. I think we can use the Crysis Benchmark tool for the bench. :) (And it also saves me the trouble of having to put my E8400 back in for a bench) Clock your E8400 as high as you can and then use the tool to obtain a result similar to this one:

http://i235.photobucket.com/albums/ee60/cuongvietpham/PC/CrysisBenchDX10.jpg

And we'll see. :) Oh, and the res should be either 1024 x 768 or 1280 x 1024 so our GPUs won't limit the bench... much. :)


It should be fun. I have a 3870 so you'll have a little advantage with that, but it'll be interesting to see how much extra horse power the e8000's need to compete. I'll run some benches and then post them.

[RunawayPrisoner]

I'm using the 9600GT then. 9600GT and 3870 are pretty much equivalent, right? :)

[3lfk1ng]

Let try to eliminate as many variables as possible...

To post your competition results your post should include:

1 test with the following:
GPUZ,
CPUZ -cpu
CPUZ -mainboard
CPUZ -memory
Crysis Benchmark at 1024x768 (aa off) (4 runs, 1st run doesn't count as it is typically filling memory)

1 test with the following:
GPUZ,
CPUZ -cpu
CPUZ -mainboard
CPUZ -memory
Crysis Benchmark at 1280x1024 (aa off) (4 runs, 1st run doesn't count as it is typically filling memory)

Just a suggestion

[Borgod]

I sold my QX6700 for an e8400 and it made my QX6700 look like a 486DX2 :D

I'll quote myself from a week or two ago on another forum:

Medium setting, 0xAA

8800GTX SLI
4gig RAM
Windows XP
1920x1200

QX6700 @ 3.2Mhz
169.28 drivers
Crysis v1.1

e8400 @ 4Mhz
169.44 drivers
Crysis v1.2

http://i4.photobucket.com/albums/y106/Terebro/PC%20misc/crybenching.jpg

I'm perfectly aware this is not a clock-for-clock comparison, but for the average overclocker these is a pretty good comparison.

[3lfk1ng]

3.2mhz...
4mhz...
:)

[RunawayPrisoner]

Well, I think this will do just fine:

One run at 1024 x 768 DX9 all High settings and with the Assault_Harbor timedemo
One run at 1024 x 768 DX10 all Very High settings and also with the Assault_Harbor timedemo.

A screenshot of CPU-Z showing which frequency the CPU is running at... and... that'll be it. Add an additional GPU-Z window to show off the GPU if you like. :) I choose 1024 x 768 because it'll stress the GPU just right to leave some room for the CPU. Anything lower than 1024 x 768 just doesn't run much faster... and anything higher than that uses way too much VRAM at Very High settings. I already have my scores ready (in fact, there's the Very High scores up there) and so... I'll post the rest when you're ready. :)

P.S.: And if we're tie, that means Crysis needs a better GPU. :D

Here are the 8800GTS G92 scores, by the way:

DX9:
http://i235.photobucket.com/albums/ee60/cuongvietpham/PC/CrysisBenchDX988gts.jpg

DX10:
http://i235.photobucket.com/albums/ee60/cuongvietpham/PC/CrysisBench88gts.jpg

I know. It totally dominates even on Very High. :p

[PhilDoc]

I'm using the 9600GT then. 9600GT and 3870 are pretty much equivalent, right? :)


I only looked at one review, tweektown. The 9600gt was 4 fps quicker than the 3870 in crysis at 1280 x 1024 and medium settings. Didn't look at the rest of the games. The GTS was 12 fps higher at same. Anyway, we'll have to wait for the final result 'til at least Friday. The memory stick that was dying on me finally said goodbye today. I tried 1 stick, but it was way too slow. I have some team group comming in on Friday, but it doesn't have heat spreaders. I'll try to run them anyway. I did get a couple of runs in. This was at 4.25G and 864/1242 on the card. Concidering the difference in cards that may be about equal, but I'll run some more in a couple of days.

[RunawayPrisoner]

Ouch... that was indeed slow. But it came close enough. :) What concerns me is that the results are almost similar except for the Min FPS. Hmm... can you try setting everything but Physics Quality to Low? And then set Physics Quality to Very High? I'm sure that your RAM won't impact anything there. In fact, I'll pull one stick out and run single channel as well. Let's do it! :)

[PhilDoc]

Ouch... that was indeed slow. But it came close enough. :) What concerns me is that the results are almost similar except for the Min FPS. Hmm... can you try setting everything but Physics Quality to Low? And then set Physics Quality to Very High? I'm sure that your RAM won't impact anything there. In fact, I'll pull one stick out and run single channel as well. Let's do it! :)


Lol, it was only 1.5 behind and you have 4 cores and a quicker card in crysis, lol. It did take .45G though. So I'd say Crysis really benifits from the Quad. Nice results with the GTS.

It's way to painful to watch and its bedtime here. Maybe tomorrow, I'll try your settings and we'll compare. Have a good one.


Edit: I'll also try the new hot fix for studdering, that may help the low fps.

[RunawayPrisoner]

Ahh... okay. :p And I have something to show:

Really really REALLY NEVER EVER try this at home: Running Crysis in Windows Vista with only 1GB of RAM. Why? :p:

http://i235.photobucket.com/albums/ee60/cuongvietpham/PC/Crysis1GVista.jpg

It stutters like there is no tomorrow. I'll see what happens when I set my RAM sticks to single-channel instead. Running Crysis even at Low... with 1GB of RAM is just plain suicidal. It was like a slide show fest...

[Borgod]

3.2mhz...
4mhz...
:)

Yep... both are as stable as eachother.

[3lfk1ng]

I was trying to point out "mhz"... when they should be "ghz" I guess you didn't catch that.

[Borgod]

Edit: you know what I meant :)

[PhilDoc]

Ahh... okay. :p And I have something to show:

Really really REALLY NEVER EVER try this at home: Running Crysis in Windows Vista with only 1GB of RAM. Why? :p:

http://i235.photobucket.com/albums/ee60/cuongvietpham/PC/Crysis1GVista.jpg

It stutters like there is no tomorrow. I'll see what happens when I set my RAM sticks to single-channel instead. Running Crysis even at Low... with 1GB of RAM is just plain suicidal. It was like a slide show fest...

Painful isn't it. You should try it on very high, lol. I let it go for a few frames, then had to stop it. I just couldn't stand it any longer. We'll play some more on Friday, but it looks like Borgod has answered the question for this thread. At least about crysis. The good news, my X3350 is out for delivery, too bad my ram is still in Sacremento, lol. I'll also try to run some with the quad vs the e8400, just to eliminate the video card.

[RunawayPrisoner]

Yeah... go for it. ;) I was about to do the same thing... had it not been for the Big Typhoon being insanely hard to remove from the board...

[PhilDoc]

Yeah... go for it. ;) I was about to do the same thing... had it not been for the Big Typhoon being insanely hard to remove from the board...

LOL, I know the feeling still no quad, but I did try my memory again and for what ever reason, maybe just cooled down, anyway, did get much out of, 900, but did some runs. Hopefully, its not my cpu. Also, it seems I was wrong about the card clocks on the other run. I thought it was at the bios default, but forgot I'd turned it down earlier. Anyway, here you go 4.05G and what amounts to 877/1323 on the card. A little better, but still really low on the min fps.

[RunawayPrisoner]

Just confirmed that the min FPS is something related to your graphics card. So don't worry. :) But here's the ultimate sentence to this one question:

Here's an E8400 @ 3.80GHz running a 8800GTS G92:
http://i235.photobucket.com/albums/ee60/cuongvietpham/PC/CrysisBenchE8400.jpg

Here's an X3350 @ 3.80GHz running the same 8800GTS G92... with quite a tad slower mem clock:
http://i235.photobucket.com/albums/ee60/cuongvietpham/PC/CrysisBench88gts.jpg

And... there you have it. The game runs about 1fps faster on average on the X3350. You may not think that that's much, but it's critical in Crysis... especially when the game is running in Very High.

Verdict: At the same clock speed, quad-core processors own in Crysis. :up: I also noticed quite a significant amount of lag with the E8400 compared to the X3350, too... even on the second run.

Edit: And here's run 0 results for E8400 if you are interested:

!TimeDemo Run 0 Finished.
Play Time: 134.33s, Average FPS: 30.52
Min FPS: 19.72 at frame 693, Max FPS: 47.64 at frame 1361
Average Tri/Sec: 15045237, Tri/Frame: 492947

Run 0 is usually where the map is still being streamed into memory, so I don't usually use run 0 as the verdict, but if I ever do, then run 0 on E8400 is still... slower than run 0 on the X3350 despite the E8400 having higher mem clocks (which speeds up the game quite a tad bit).

[PhilDoc]

I think most already conceded that the quad would be quicker clock for clock. The thing about an e8400 is that most of the time you'll get a higher OC and it looks like from your test, would be the best choice for Crysis.

[RunawayPrisoner]

Yeah, I guess... if it's not that E8400 suffers from degradation and that the voltage requirement to stay stable may vary a lot, and sometimes, the average user can only get to 3.6GHz or 4.0GHz stable on the E8400.

[Grnfinger]

Yeah, I guess... if it's not that E8400 suffers from degradation and that the voltage requirement to stay stable may vary a lot, and sometimes, the average user can only get to 3.6GHz or 4.0GHz stable on the E8400.

I dont know the 8400 is a good choice for great clocks and low price.
Beats selling a kidney to get a QX

http://i237.photobucket.com/albums/ff267/Grnfinger/4300prime1.jpg

http://i237.photobucket.com/albums/ff267/Grnfinger/test2-2.jpg

http://i237.photobucket.com/albums/ff267/Grnfinger/4500_keep.jpg

priming 4500 now

[PhilDoc]

Yeah, I guess... if it's not that E8400 suffers from degradation and that the voltage requirement to stay stable may vary a lot, and sometimes, the average user can only get to 3.6GHz or 4.0GHz stable on the E8400.

I would say the vast majority is more like 4G and that's because many wont give it a lot of volts. I've primed mine at 4400 on air and less then 1.4v. Very few have suffered from degradation and your X3350 will have the same risk. So you may want to drop your PLL a little. Its been know to harm cpus, even 65nm ones.

[RunawayPrisoner]

My PLL is running at stock, so is my FSB Termination volt. The only thing that needs to be high enough for the mainboard to handle such FSB is the Northbridge voltage. All of the other values are only increased when I need results from 485MHz or above.

And by the way, I am kind of laying off priming now. If I want to know that my system is stable enough, I'd run that Crysis benchmark over and over again... overnight, at the highest settings possible. :) Sincerely, even a prime stable system can fail after 2 or 5 runs when Crysis is at its max. I'm only priming now... when I want to find out if my cooler is good enough. By the way, here's the 9600GT with DX10 run.

http://i235.photobucket.com/albums/ee60/cuongvietpham/PC/CrysisDX10.jpg

Seems like it beats your 3870... still, even when the 3870 is overclocked to that extend. I'm guessing that even at 4G, E8400 wouldn't be enough for Crysis... but hey, maybe that's just something... hmm... I'll put the E8400 in again tomorrow for another run at it.

[[XC] NetburstXE]

I'd get the QX6700. That is an insane deal you have. It should OC to 3.4 quite easily. Although the E8400 will beat it in games only optimized for 1-2 cores (due to a higher OC potential), the QX6700 is more future-proof. In the future, there will be more multi-threaded apps where the quad core CPU will really shine. Also, for folding 4 cores will be simply amazing.

[PhilDoc]

My PLL is running at stock, so is my FSB Termination volt. The only thing that needs to be high enough for the mainboard to handle such FSB is the Northbridge voltage. All of the other values are only increased when I need results from 485MHz or above.

And by the way, I am kind of laying off priming now. If I want to know that my system is stable enough, I'd run that Crysis benchmark over and over again... overnight, at the highest settings possible. :) Sincerely, even a prime stable system can fail after 2 or 5 runs when Crysis is at its max. I'm only priming now... when I want to find out if my cooler is good enough. By the way, here's the 9600GT with DX10 run.

http://i235.photobucket.com/albums/ee60/cuongvietpham/PC/CrysisDX10.jpg

Seems like it beats your 3870... still, even when the 3870 is overclocked to that extend. I'm guessing that even at 4G, E8400 wouldn't be enough for Crysis... but hey, maybe that's just something... hmm... I'll put the E8400 in again tomorrow for another run at it.


That's not high for a 3870. Mine will do 891/1400. Many will do more. Mine is probably average. Good, irrc you said your PLL was 1.8v in the X3350 thread.

[WittieM]

Well it looks like I will be going with the qx6700 then. If someone doesn't mind taking a look at the planned components and let me know if anything should be revised, added or deleted.

eVGA 780i mb
QX6700
2GB Corsair Dominator PC8500 (will move to 4gb later)
Graphics cards TBD? Between 8800GTX, 8800GT/S 512mb, 9800GX2 (I still have the 2x 790GTXs for temp)
PC Power & Cooling PS. Not sure on which one yet.

[RunawayPrisoner]

That's not high for a 3870. Mine will do 891/1400. Many will do more. Mine is probably average. Good, irrc you said your PLL was 1.8v in the X3350 thread.

It was at that level to reach 500MHz. At 475MHz, I can let the motherboard take care of everything else but the NB volt, which I have to set manually to 1.5 -1.55v for the whole rig to be stable.