[psolord]

Hello Vr-Zone friends. First of all I would like to extend my best wishes for a Happy New Year to you and your families.

I've recently acquired a Core i7-860 and I decided to do some gaming tests on it and compare the results with those of my Q9550. I used my ATI Radeon 5850 clocked at 1000Mhz core and 1250Mhz RAM for both systems, bar a couple of benchmarks that proved to be more difficult and I was forced to reduce to 950-975Mhz for both systems anyway.

The rest of the systems specs are as follows:

Intel Core i7-860 @4.0Ghz
Asus Maximus III Formula
Geil 4GB DDR3-2133 modules
Enermax Galaxy 850W PSU
Windows 7 X64

Intel Q9550 @4.0GHz
Asus P5Q Deluxe
OCZ 4GB DDR2-1000 modules
Nexus 850W PSU
Windows 7 X64

As stated, both CPUs were overclocked to 4.0Ghz and I used 43 gaming benchmarks to have a better picture. The ugly part is that in six benchmarks, the Core i7-860 was overclocked to 4.2Ghz. These were my first benchmarks, but then I decided that the extra voltage was not worth it and dropped to 4.0Ghz for the rest of the benchmarks.These six benchmarks were GTA IV, ARMA 2, SERIOUS SAM HD, DIRT 2, AVATAR, UNIGINE HEAVEN. 200Mhz can hardy make a difference anyway!

One more ugly thing is that this test spanned across three different driver sets, namely Catalyst 9.10, 9.11 and 9.12. Catalyst 9.11 was used in the six aforementioned 4.2Ghz Core i7 benchmarks as well as GTA IV, SERIOUS SAM HD, DAWN OF WAR II and DIRT 2 for the Q9550 system. The rest are all with Catalyst 9.12 for the Core i7 and Catalyst 9.10 for the Q9550.

Yes I know that you don't do gaming CPU comparison with different gfx card driver sets, but hey, it is widely known that Catalyst 9.10 through 9.12 brought no significant performance differences, so that's why I decided to publish it anyway. In any case, if a system is at a disadvantage here, that's the Q9550 one, since it uses the older drivers and as you will see, it gave quite a battle. I could even spoil it for you, by saying that these two systems presented very small difference overall, but I won't! Hey wait a minute!


The benchmark results will be presented threefold, meaning, there is one plain text table, of which the framerates/results are actual links to their respective 1080P video in Youtube, there is an excel spreadsheet with which I calculated the percentage differences of the two CPUs (sorted from the biggest performance difference to the smallest) and there are also a few good ol' bar graphs.

Enough talk. Enjoy the benchmarks.


First is the text table. Don't forget to click on the framerate result if you want to watch the respective video.

For all the following benchmarks, an ATI RADEON @1.0Ghz has been used.

The resolution is 1920X1080. Max available Anisotropic Filtering has been used in all cases.

------------------------------------------------------------------------------------
|Game - Benchmark ......................|.Quality.|......ATI.RADEON.5850@1.0GHz....|
|#######################################|#########|CORE.I7-860@4.0GHz|.Q9550@4.0GHz|
|#######################################|#########|################################|
|----------------------------------------------------------------------------------|
|STALKER.CLEAR.SKY.DX.10.SUNRAYS.ON.....|...4XAA..|.....40,637...|.....40,530......|
|----------------------------------------------------------------------------------|
|UNIGINE.TROPICS.DX10...................|...4XAA..|.....61,800...|.....60,900......|
|----------------------------------------------------------------------------------|
|CALL.OF.JUAREZ.DX10.BENCH..............|...4XAA..|.....67,800...|.....66,300......|
|----------------------------------------------------------------------------------|
|UNIGINE.HEAVEN.BENCHMARK.DX11......... |...NOAA..|.....40,500...|.....41,200......|
|----------------------------------------------------------------------------------|
|CRYSIS.DX10.VERY.HIGH..................|...NOAA..|.....38,565...|.....38,160......|
|----------------------------------------------------------------------------------|
|RED.FACTION.GUERRILLA.DX10.VERY.HIGH...|...8XAA..|.....40,235...|.....40,975......|
|----------------------------------------------------------------------------------|
|CRYSIS.WARHEAD.AMBUSH.ENTHUSIAST.DX10..|...NOAA..|.....38,350...|.....38,040......|
|----------------------------------------------------------------------------------|
|COMPANY.OF.HEROES.DX10.BENCH...........|...4XAA..|....125,700...|....124,800......|
|----------------------------------------------------------------------------------|
|CRYSIS.DX10.HIGH.......................|...NOAA..|.....62,925...|.....61,160......|
|----------------------------------------------------------------------------------|
|FAR.CRY.2.DX10.DEMO.PLAYBACK...........|...4XAA..|.....67,260...|.....65,500......|
|----------------------------------------------------------------------------------|
|TRINE..................................|...4XAA..|.....66,455...|.....66,603......|
|----------------------------------------------------------------------------------|
|CRYSIS.WARHEAD.AVALANCHE.DX10.GAMER....|...NOAA..|.....60,820...|.....58,650......|
|----------------------------------------------------------------------------------|
|CALL.OF.DUTY.V.WORLD.AT.WAR............|...4XAA..|.....95,873...|.....98,541......|
|----------------------------------------------------------------------------------|
|RACE.DRIVER.GRID.......................|...4XAA..|....120,432...|....108,988......|
|----------------------------------------------------------------------------------|
|ARMA.II.HIGH.PRESET....................|..NORMAL.|.....44,000...|.....58,000......|
|----------------------------------------------------------------------------------|
|BATMAN.ARKHAM.ASYLUM.NOPHYSX...........|...NOAA..|....167,000...|....170,000......|
|----------------------------------------------------------------------------------|
|FAR.CRY.2.DX10.LONG.RANCH..............|...4XAA..|.....76,430...|.....72,820......|
|----------------------------------------------------------------------------------|
|MASS.EFFECT............................|...NOAA..|....119,766...|....106,684......|
|----------------------------------------------------------------------------------|
|WORLD.IN.CONFLICT.ALL.MAXED............|...4XAA..|.....57,000...|.....56,000......|
|----------------------------------------------------------------------------------|
|DEVIL.MAY.CRY.IV.DX10.SCENE.1..........|...4XAA..|....163,040...|....144,890......|
|----------------------------------------------------------------------------------|
|DEVIL.MAY.CRY.IV.DX10.SCENE.2..........|...4XAA..|....129,430...|....114,210......|
|----------------------------------------------------------------------------------|
|DEVIL.MAY.CRY.IV.DX10.SCENE.3..........|...4XAA..|....206,310...|....173,060......|
|----------------------------------------------------------------------------------|
|DEVIL.MAY.CRY.IV.DX10.SCENE.4..........|...4XAA..|....121,330...|....107,470......|
|----------------------------------------------------------------------------------|
|ANNO.1404..............................|...4XAA..|.....69,246...|.....61,284......| <- It seems I haven't uploaded a video here, lol
|----------------------------------------------------------------------------------|
|BIOSHOCK.DX10..........................|...NOAA..|....163,547...|....153,302......|
|----------------------------------------------------------------------------------|
|CALL.OF.DUTY.IV........................|...4XAA..|....151,420...|....164,093......|
|----------------------------------------------------------------------------------|
|STREET.FIGHTER.IV......................|...4XAA..|....113,780...|....117,700......|
|----------------------------------------------------------------------------------|
|HALF.LIFE.2.EPISODE.2..................|...4XAA..|....200,664...|....188,921......|
|----------------------------------------------------------------------------------|
|RES.EVIL.5.DX10.BENCH..................|...4XAA..|....110,400...|....102,400......|
|----------------------------------------------------------------------------------|
|OBLIVION...............................|...NOAA..|....116,730...|.....81,515......|
|----------------------------------------------------------------------------------|
|UNREAL.TOURNAMENT.III..................|...NOAA..|....154,000...|....149,000......|
|----------------------------------------------------------------------------------|
|ENEMY.TERRITORY.QUAKE.WARS.............|...4XAA..|....167,000...|....134,000......|
|----------------------------------------------------------------------------------|
|CELLFACTOR.REVOLUTION..................|...NOAA..|.....75,581...|.....47,533......|
|----------------------------------------------------------------------------------|
|CLIVE.BARKER'S.JERICHO.................|...4XAA..|.....67,420...|.....67,276......|
|----------------------------------------------------------------------------------|
|FALLOUT.3.INDOORS.DX9.ULTRA............|...4XAA..|.....95.661...|.....83,631......|
|----------------------------------------------------------------------------------|
|FALLOUT.3.OUTDOORS.DX9.ULTRA...........|...4XAA..|.....79.014...|.....66,596......|
|----------------------------------------------------------------------------------|
|ASSASSIN'S.CREED.DX10..................|...NOAA..|.....76,340...|.....68,734......|
|----------------------------------------------------------------------------------|
|BIONIC.COMMANDO........................|...NOAA..|....116,935...|....116,546......|
|----------------------------------------------------------------------------------|
|FUEL.DX9.............................. |...4XAA..|.....74,011...|.....63,407......|
|----------------------------------------------------------------------------------|
|TRACKMANIA.FOREVER.UNITED..............|...4XAA..|.....95,500...|.....82,600......|
|----------------------------------------------------------------------------------|
|GRAND.THEFT.AUTO.IV.DX9............... |...NOAA..|.....45,842...|.....36,083......|
|----------------------------------------------------------------------------------|
|COLIN.MC.RAE.DIRT.2.DX11.............. |...4XAA..|.....63,900...|.....62,300......|
|----------------------------------------------------------------------------------|
|JAMES.CAMERON.AVATAR.DX.10............ |...8XAA..|.....59,933...|.....55,274......|
|----------------------------------------------------------------------------------|
|WARHAMMER.40000.DAWN.OF.WAR.II.ULTRA.. |..AA.ON..|.....77,410...|.....64,500......|
|----------------------------------------------------------------------------------|
|TOM.CLANCY'S.HAWX.DX10.................|...4XAA..|.....75,000...|.....73,000......|
|----------------------------------------------------------------------------------|
|SERIOUS.SAM.HD.........................|...NOAA..|....166,655...|....161,944......|
|----------------------------------------------------------------------------------|
|CALL.OF.DUTY.MODERN.WARFARE 2..........|...4XAA..|....104,303...|....105,316......|
|----------------------------------------------------------------------------------|



The excel sum up follows hosted in two imagehosters!

Imagehosting.gr


Imageshack.us








And last but not least the graphs that we all love!











Hope somebody finds this useful.

ATI 5850 Crossfire tests will follow, but by the time they will be published the Catalyst 9.1 will be out! Oh well..!

Cheers!

Addition 1 - Non gaming related videos, of the Core i7 vs the Q9550

YouTube - WinRar 3.80 Benchmark - Intel Q9550 @4.0Ghz vs Intel Core i7 860 @4.2Ghz

and also the same test with winrar 3.91 that added hyperthreading support (sorry no Q9550 here)

YouTube - Intel Q9550 @4Ghz vs Intel Core i7 860 @4.2Ghz Everest Benchmarks

YouTube - Intel Q9550 @4.0Ghz vs Intel Core i7 860 @4.0Ghz CINEBENCH 10 Benchmark

YouTube - Intel Q9550 @4.0Ghz vs Intel Core i7 860 @4.1Ghz Sisoft Sandra Proffessional 2010

Share|

[9VARZ]

That's an epic test!

[MrBungle]

I'm impressed. Great work. Now I wouldn't feel bad about upgrading my E8400 to a Q9550. I've been holding back all this time because I thought it would be a waste to buy a CPU that wasn't Core i7 or i5. Now I see the 9550 is more than fast enough.

[Soviet_Rebel]

Holy CRAP! That's an impressive awesome work! I find myself, for the second time in a short period of time, giving kudos to you!

I agree with MrBungle, Core2Quad processors (clocked at the right speed) are still very fast. I7's cannot unleash their true power in single GPU configuration; try that monster in crossfire and you will see a HUGE improvement!

Keep up the good work psolord, i appreciate it.

[haylui]

Originally Posted by MrBungle

I'm impressed. Great work. Now I wouldn't feel bad about upgrading my E8400 to a Q9550. I've been holding back all this time because I thought it would be a waste to buy a CPU that wasn't Core i7 or i5. Now I see the 9550 is more than fast enough.

lolz
i5 and i7 are faster in video encoding and some multimedia processor jobs
not entirely faster in all arena.
if not Phenom II could only be competing with Celeron only
Let's relocate your resources into graphic processor rather than CPU if you're into gaming
i5,i7 and i9 won't give you a huge boost in gaming performance!!!

[Soviet_Rebel]

i5 and i7's are faster even in gaming. In psolord's scenario the gpu is bottlenecking the processors; as i said before to unleash the true power of those monsters you must have a multi-GPU configuration. With a high level multi-gpu configuration you will see a huge boost in gaming performance.

Since MrBungle's CPU is fast enough for gaming he would for sure do better upgrading his GPU (anyway the 4890 is still a monster), but he would do much better upgrading to i7 + 5970 crossfire...
...
...
...

[9VARZ]

Originally Posted by haylui

lolz
i5 and i7 are faster in video encoding and some multimedia processor jobs
not entirely faster in all arena.
if not Phenom II could only be competing with Celeron only
Let's relocate your resources into graphic processor rather than CPU if you're into gaming
i5,i7 and i9 won't give you a huge boost in gaming performance!!!

I5/i7 are also faster for professional apps like MATLAB and such. But you are right, even those are not everyday use de. Everyday use PC, AMD enough liao.

[Magnuz]

Great work of love!
More extensive than even reviews from reputable sites.

Gives me more reason to hang on to the Quad cores!

Appreciate it.

[DaGamer!!!]

Why do you guys think that when I'd gotten my 2x HD5870's, I did not bother with upgrading my CPU/mobo/RAM. My X9650 @3.9ghz is fast enough for me, sure my FR and benchmark score might go up with an i7 920/X58 combo, but as long as my CPU can keep FR at more than playable level, should I even bother?

Someone quoted a 'huge performance difference' with multi-GPU i5/i7 combo vs multi-GPU C2Q, while I believe that yes, in terms of absolute numbers that the i5/i7 combo can crunch, it'd win over the C2Q. But what does that mean? If the Q9550 + HD5970 post 70fps in one game, while the i5/i7 + HD5970 post 80fps in the same game, would that matter to me? Both would be smooth, so it'd add nothing to my gaming experience. I'd rather spend the money on a possible i7/X58/DRR3 upgrade on, say, 2x 24" monitors for tri-monitor EyeFinity gaming......and guess what? I did exactly that!

[catacomb]

nice clock.. good work.. tonnes of effort

[ORiN]

Great effort from psolord again! It will be good if you can add some video encoding benchmarks.

[watzup_ken]

That is one hell lot of work running those tests. Anyway, the Core i7 either performs on par or much better in those test, depending on the tests. I suppose if an application really puts a lot of load on all 4 cores at one go, then i7 should pull ahead. I suppose previous gen quad users like me should not worry too much on the performance difference unless you are running very taxing applications like video encoding on a frequent basis.

[power666]

Impressive testing. Now for various loose end questions.

How many of the games tested scaled across all four cores on the Q9550? How many could handle the 8 virtual cores on the i7-860? Did you keep track if Turbo was kicking in the i7-860? That could boost the clock speed to 4.2 Ghz or 4.4 Ghz in some situations if your cooling was adequate. Speaking of, what do you use to cool your rigs?

I suspect that the bulk of the increase in CPU performance is due to higher memory bandwidth and the latency improvement via the on-die memory controller on the i7-860. This is something that'd affect all games in a positive manner.

Also why do the Excel graphs have 5850 OC vs. GTX260 OC in the upper right hand corner?

[psolord]

Thanks everyone for their kind comments and input.

Yes the Q9550 is still an awesome cpu. I do have the feeling though that this test does not show what the Core i7 is truelly capable of! Soviet Rebels is pretty much spot on about what the Core i7 can do in multi gpu. My early Crossfire results have shown very interesting things.

For example ~100fps in Far Cry 2 for the Q9550 while I get ~120fps on the same test with the core i7! That's a solid 20% performance gain from the cpu alone! It is also noteworthy that the scaling of this game is great. From 67fps to 120fps with the second card is just plain awesome!

Dagamer has a very correct perpsective though. Since we are talking about games here, it is much more preferable to install a second graphics card (or just a better one) than a new cpu/mobo/ram. To speak with money ammount, the core i7/mobo/ram stuff cost about 700 euros, while the second 5850 cost just 220 euros. It is obviously a much more reasonable choice for what you get back for your money. In any case I built a new rig not upgraded an old one and I am a guy that loves his benchmarks (obviously lol). This test and the crossfire test will hopefully put things in better perspective for some people.

@Watzup ken, I have recorded some other tests that I will add in my first post, but no video encoding tests unfortunately. As you will see from these other tests, the i7 is quite a little number cruncher! My Cellfactor benchmark is a good indication of that too!

@Power666
Lol I am stupid, that's why you see that 5850 OC cs GTX 260 up there. Let me fix that!

I didn't do a specific study about cpu usage since I know from experience that most games are dual threaded at best anyway! I can say for sure that the 8 virtual threads are fast asleed in most cases, enen in applications. The only application that put 100% cpu load was Xilisoft Video Converter with 8 simultaneous video encodings running. Even with 4 encodings, the task manager shows only 50% cpu load which is obviously wrong. I think either Intel or Microsoft are cheacting in cpu measurement. I mean there is no way that by not using the spare Hyperthreading you have 50% more. It would be a 25% more at best (right?)!

I believe that, whatever better performance comes in games, is mainly due to the per thread better performance of the core i7, since it sports better/newer/more efficient architecture and of course higher bandwidth and the on dies PCIe controller. Again I will mention Cellfactor though. The i7 chews through all that physics calculations and I think that the extra bandwidth is not doing much here. Oblivion is another case altogether though. Bandwidth plays a greater role there.

[power666]

Originally Posted by psolord

I didn't do a specific study about cpu usage since I know from experience that most games are dual threaded at best anyway! I can say for sure that the 8 virtual threads are fast asleed in most cases, enen in applications. The only application that put 100% cpu load was Xilisoft Video Converter with 8 simultaneous video encodings running. Even with 4 encodings, the task manager shows only 50% cpu load which is obviously wrong. I think either Intel or Microsoft are cheacting in cpu measurement. I mean there is no way that by not using the spare Hyperthreading you have 50% more. It would be a 25% more at best (right?)!

I've seen several cases where a well threaded application can only reach about 50% load per virtual CPU. It'll try and saturate all the threads to 100% theoretical load regardless if they are a virtual CPU core or not. The main benefits for Hypthreading is in a mixed environment where there is a mix of threads utilizing different CPU resources. Several instances of the same thread can easily create bottlenecks in the system as they compete for the same number of limited resources.

Does the CPU load jump to 100% on four cores when Hyperthreading is disabled?