gtx260 216 OC vs. gtx465 OC benchmarks

[tviceman]

I upgraded, and to my own surprise I went with a GPU I thought I'd never even consider purchasing - the gtx465. However this is no ordinary gtx465; it's the MSI twin frozr II edition. As many of you know, this is among the best (if not THE BEST) coolers that you can get preassembled on a graphics card. And as I will explain in a bit, this cooler does an outstanding job of taming this seriously cut down version Fermi. It came with a copy of Mafia II (which I already own) and a $60 mail in rebate. It cost me $212 shipped, so after I sell the extra copy of Mafia II ($30ish) and I get the MIR, the total cost to me will be $122. NOT BAD AT ALL! I had my heart set on getting the ASUS directCU TOP gtx460 1gb that was on sale at amazon for $210 (and briefly $175) but after ordering it on Aug. 8th and still not having received it, I cracked the whip on MSI card. I still haven't canceled my amazon order yet... I'll decide what to do about that later.

So I'll get straight to it. Running at 1.075 volts, this card is 100% stable with clocks of 825mhz core (up from 607mhz) and 3500mhz memory (up from 3200mhz). This represents a 36% increase in core/shader speeds and a little over 9% for memory. I'd like to get the memory up to 3600, but I haven't tested it out at that speed yet and even if it can't hit 3600 reliably I'm happy with how it is now. The card's idle temperature is in the mid 30's, and the GPU tops out at 77 degrees celsius in MSI Kombustor (Furmark) when at 60% fan speed. 60% fan speed is audible if your speakers are not at any appreciable sound level. In other words, if you can clearly hear the game's sound effects and music, it will entirely drown out the GPU fan noise. Anything below 50% is completely inaudible. I have my fan set to operate between 40-80% based on the temperature of the GPU, but unless something terrible goes wrong it will simply never go above 60% fan speed.

So now the benchmarks... All tests are ran at 1680x1050 max in-game settings with AA and AF on where indicated.

Crysis custom config.sys DX9 mix between high/very high
gtx260 avg: 41.7, min: 30.8, max: 52.9
gtx465 avg: 68.4, min: 24.7, max: 78.2....................Average frame rate up 64%, minimum frame rate DOWN 20%

Crysis DX10 enthusiast
gtx260 avg: 26.9, min: 17.7, max: 32.5
gtx465 avg: 37.0, min: 15.0, max: 41.7....................Average frame rate up 38%, minimum frame rate DOWN 15%

Crysis Warhead DX10 Enthusiast
gtx260 avg: 33.1, min: 23.2, max: 41.0
gtx465 avg: 43.0, min: 26.7, max: 50.9....................Average frame rate up 31%, minimum frame rate up 15%

Batman: Arkham Asylum 4X AA, 16X AF, physx high
gtx260 avg: 43, min: 23, max: 65
gtx465 avg: 66, min: 35, max: 109...........................Average frame rate up 53%, minimum frame rate up 52%

Far Cry 2 Small Ranch 4XAA 16X AF
gtx260 avg: 53.7, min: 41.0, max: 78.9
gtx465 avg: 62.1, min: 41.2, max: 96.2....................Average frame rate up 16%, minimum frame rate essentially unchanged

Far Cry 2 Action Scene 4XAA
gtx260 avg: 38.0, min: 32.3, max: 48.6
gtx465 avg: 36.8, min: 30.7, max: 45.2....................Average frame rate DOWN 3%, minimum frame rate DOWN 5%

Darkest Of Days 4XAA, 8XAF, physx HIGH
gtx260 avg: 25.0
gtx465 avg: 35.6..................................................Average frame rate up 42%

Cryostasis Tech Demo, physx on
gtx260 avg: 42.2
gtx465 avg: 52.8..................................................Average frame rate up 25%

Just cause 2 4XAA 16XAF, bokah filter OFF
gtx260 avg: 38.3
gtx465 avg: 63.1..................................................Average frame rate up 65%

Just Cause 2 4XAA 16XAF bokah filter ON
gtx260 avg: 34.8
gtx465 avg: 58.4..................................................Average frame rate up 68%

Overall those are some pretty significant performance improvements. Nearly everything I've tested so far experienced substantial increases in average frame rates. Interesting results are observed in Far Cry 2 and the original Crysis, though. The minimums are down and the average was down in the Far Cry 2 action playback benchmark. This is telling me that the gtx465, while capable of processing larger amounts of data than the gtx260 216, is not quite as efficient as the gtx260. It is my best guess that the gtx260's more texturing units (64 vs. 44 in the gtx465) allows the GPU to process certain scenarios more efficiently, evident when the CPU is either limiting the performance of the game or the GPU is using an extremely heavy amount of textures.

***the gtx260 216 was running at 675mhz core, 1450mhz shader, and 2150mhz effective memory speeds (stock speeds are 576 core, 1242 shader, and 2000 mem)

My system specs are as follows:

Core2duo e8400 @ 4.05ghz
Gigabyte EP45-UD3P Motherboard
4 gigs 1066 DDR2 RAM cas 5
Win 7 x64
1680x1050 native resolution

 

[RussianSensation]

Nice results, but what resolution? I expect GTX4xx series to pull away from GTX2xx series in the next 6 months or more modern games come out (see how much faster it is in Just Cause 2).

Also, if you have Dirt 2, Metro 2033 or STALKER:CoP, the 465 overclocked should be a lot faster than the GTX260 as well.

 

[tigersty1e]

what are stock clocks of the 260?

and what are the rest of your system specs?

 

[tviceman]

Nice results, but what resolution? I expect GTX4xx series to pull away from GTX2xx series in the next 6 months or more modern games come out (see how much faster it is in Just Cause 2).

Also, if you have Dirt 2, Metro 2033 or STALKER:CoP, the 465 overclocked should be a lot faster than the GTX260 as well.

Sorry I left out important specific information.

My system specs are as follows:

Core2duo e8400 @ 4.05ghz
Gigabyte EP45-UD3P Motherboard
4 gigs 1066 DDR2 RAM cas 5
Win 7 x64

All benchmarks were ran at my monitors native resolution: 1680x1050. I own Metro2033, but unfortunately did not get a chance to run the game's built in benchmark with my gtx260 before giving it to my brother. I have a 8800GTS 512 (aka 9800gtx) coming to me in the mail, and I am going to run a gamut of physx performance tests involving no dedicated card and using the 8800GTS 512 as a dedicated physx card. The games I will test are Mafia II, Cryostasis, Metro2033, Darkest of Days, and Batman: AA.

 

[tviceman]

what are stock clocks of the 260?

and what are the rest of your system specs?

Sorry forgot to include that. The gtx260 216 I had was running at 675mhz core, 1450mhz shader, and 2150mhz effective memory speeds vs. stock speeds are 576 core, 1242 shader, and 2000 mem.

 

[Skurge]

A lot of people will say that wasn't a worthwhile upgrade.

But what do you think? Do you think it was worth it? From the speed increases I would say it was.

 

[tviceman]

A lot of people will say that wasn't a worthwhile upgrade.

But what do you think? Do you think it was worth it? From the speed increases I would say it was.

I think after selling Mafia II and getting the MIR, $122 this will be well worth this upgrade. And I don't play Far Cry 2. I own the game, and it has a great built in benchmark, but I think about 8 hours was all I could stand. But still factoring that game into my results, getting > 40% performance in half the benchmarks I ran indicates to me that, yeah, for $122 it was worth it.

 

[toyota]

sorry but some of those numbers just don't make sense. I most certainly did not have that big of a jump going from a 192sp gtx260 to even a gtx470 using a similar cpu. overall your numbers blow away what any review shows for most of those games. heck techreport showed TWO fps difference between a gtx260 and gtx465 at 1920x1080 with 4x AA. http://techreport.com/articles.x/19242/7

 

[Flipped Gazelle]

sorry but some of those numbers just don't make sense. I most certainly did not have that big of a jump going from a 192sp gtx260 to even a gtx470 using a similar cpu. overall your numbers blow away what any review shows for most of those games. heck techreport showed TWO fps difference between a gtx260 and gtx460 at 1920x1080 with 4x AA. http://techreport.com/articles.x/19242/7

Was TR's GTX 460 overclocked? The OP's 465 is.

 

[Flipped Gazelle]

tviceman: great post. :thumbsup:

I think you did well not getting the Asus Cu - I've read enough scattered reports of issues with them that I think you did better getting the MSI.

 

[toyota]

Was TR's GTX 460 overclocked? The OP's 465 is.

I meant 465. none of their cards were overclocked. his gtx260 and gtx465 were BOTH overclocked so the massive performance gap in some games, especially Just Cause 2, doesn't jive with any official benchmarks.

 

[tviceman]

sorry but some of those numbers just don't make sense. I most certainly did not have that big of a jump going from a 192sp gtx260 to even a gtx470 using a similar cpu. overall your numbers blow away what any review shows for most of those games. heck techreport showed TWO fps difference between a gtx260 and gtx465 at 1920x1080 with 4x AA. http://techreport.com/articles.x/19242/7

LOL for some reason I knew that if you were going to comment on this thread, it would be in some negative fashion. Why would I make any of this up? Would you like screen shots to prove the benchmarks I am getting with my gtx465? Which game results do you want screen shots for?

EDIT: TR's benchmarks shows they have GPU water simulation AND the bokah filter disabled. While I'm at a lower resolution, I have both of those features enabled and I'm absolutely blowing their benchmark out of the water. I do have a 35% core overclock though.

 

[notty22]

Congrats on the purchase.
Here is a review of that card, and they tweaked it to 860 !
Here is a Batman bench where its faster than a stock gtx 470
http://www.tweaktown.com/reviews/33...overclocked_w_voltage_adjustment/index10.html
From the many things I've read, some 465's end up achieving the highest clocks because more parts are fused off. You can see when they are both at 607 the 465 is sometimes 20% slower.

 

[toyota]

LOL for some reason I knew that if you were going to comment on this thread, it would be in some negative fashion. Why would I make any of this up? Would you like screen shots to prove the benchmarks I am getting with the gtx465?

excuse me for just being honest. its nothing personal but SOME of those numbers just seem too good to be true. we have the same cpus and I saw way less of a performance jump going from even a slower 192sp gtx260 to an even faster gtx470. I am just going by MY experience and also what most sites have published. maybe that extra oc on the gtx465 helped quite a bit though.

 

[Flipped Gazelle]

I meant 465. none of their cards were overclocked. his gtx260 and gtx465 were BOTH overclocked so the massive performance gap in some games, especially Just Cause 2, doesn't jive with any official benchmarks.

But tviceman's GTX 465 OC is greater than his GTX 260 OC.

GTX 465: 607->825, 36%
GTX 260: 576->675, 17%

PS - tviceman, that's a bangin' OC on the 465!

 

[tviceman]

excuse me for just being honest. its nothing personal but SOME of those numbers just seem too good to be true. we have the same cpus and I saw way less of a performance jump going from even a slower 192sp gtx260 to an even faster gtx470. I am just going by MY experience and also what most sites have published.

What benchmark do you want me to take a screen shot for? I'll do it right now. Pick a few. I'll take screen shots of MSI afterburner too, to show you temp and fan levels throughout the bechmark.

 

[toyota]

What benchmark do you want me to take a screen shot for? I'll do it right now. Pick a few.

I don't think you are lying. I am simply just shocked that it would be that much of a difference. then again that additional oc on 465 would make it about as fast as a gtx470.

 

[tviceman]

I don't think you are lying. I am simply just shocked that it would be that much of a difference. then again that additional oc on 465 would make it about as fast as a gtx470.

EDIT: Errr I have no idea why the MSI afterburner said my core clock is at 902 on the left hand side. I didn't set it to that, I have no desire to fry my card on the second day that I own it, and the right hand clearly shows that it maxed out at 825 during the benchmark. Weird.

 

[happy medium]

Everone argues with me and says that the gtx465 is a crap gpu, Thanks for proving me right.
Nice card man. Selling the gtx260 would make it a 30$ upgrade just incredible.

 

[tviceman]

Everone argues with me and says that the gtx465 is a crap gpu, Thanks for proving me right.
Nice card man. Selling the gtx260 would make it a 30$ upgrade just incredible.

I thought it was crap too until I saw this card on sale and read the reviews.

 

[cusideabelincoln]

I don't think you are lying. I am simply just shocked that it would be that much of a difference. then again that additional oc on 465 would make it about as fast as a gtx470.

At that kind of an overclock it should be, almost unquestionably, faster than a stock 470.

sorry but some of those numbers just don't make sense. I most certainly did not have that big of a jump going from a 192sp gtx260 to even a gtx470 using a similar cpu. overall your numbers blow away what any review shows for most of those games. heck techreport showed TWO fps difference between a gtx260 and gtx465 at 1920x1080 with 4x AA. http://techreport.com/articles.x/19242/7

I don't think the jump he's seeing is too unreasonable.

http://techpowerup.com/reviews/ASUS/GeForce_GTS_450_TOP_DirectCU/28.html

The summary across all game @ 1920x1200 shows a stock GTX 465 being 21% faster than a stock GTX 260-216. Add 20% since he has a 20% higher relative overclock on his GTX 465 than he did on the GTX 260, and that would mean he shoudl be around 40% faster on average, which seems to be what tvice is getting as an average for the games he's tested (ranging from 16% to 68% improvement).

Now we can specifically look at Crysis for estimations:

At TPU the GTX 260 gets 16.8 fps and the 465 gets 20.2 fps in Crysis. If we add 17% to the GTX260 score and 36% to the GTX465 score we get 19.7 fps and 27.5 fps. The estimated GTX465-OC would now be 36% faster in Crysis than an overclocked GTX260, and this number is close to tvice's 38% improvement in DX10 mode.

I also looked at Far Cry 2 but the numbers weren't matching. I think his processor is bottlenecking the 465 in this game. If we do the same analysis at TPU:
GTX260 gets 47.4 fps (55.5 simulated 17% overclock)
GTX465 gets 64.8 fps (88.1 simulated 36% overclock)

He could be seeing a 58% increase in performance, but he only reports 16% for the benchmark. My theory that the C2D is holding him back comes from the fact that tvice reports 53.7 fps with his overclocked GTX260, which is very close to the simulated number from TPU, while his GTX 465 gets 62.1 fps, which is actually lower than the stock score from TPU. TPU is using an overclocked i7, which is known to be loved by FC2, which could be the reason for his lackluster gain.

 

[cusideabelincoln]

Everone argues with me and says that the gtx465 is a crap gpu, Thanks for proving me right.
Nice card man. Selling the gtx260 would make it a 30$ upgrade just incredible.

They were crap when the debuted at the price they debuted. Just like the HD 5830. Now both cards are much cheaper and are actually worth considering.

 

[4ghz]

They were crap when the debuted at the price they debuted. Just like the HD 5830. Now both cards are much cheaper and are actually worth considering.

Yeah a lot of cards get thrown into the "crap" pile based on initial reviews or launch prices. Worse yet many of them are never able to shake that stigma. For example the 465 uses about 10 watts more at idle and 25-40 watts more under load compared to the 460 1gb with about the same performance. Sure that's a notable difference but its not the end of the world. Yet the 460 is the current market darling and the 465 is considered a rolling pile of turd. But I guess its a good thing for people that keep their eye on the ball. A gtx 465 was a joke at $280. At $150 with a free game? Its a slamming card.

 

[Shmee]

very nice oc results! Looks like the fermis tend to OC better and better these days! Makes me wonder if I should have gotten 480's instead of 5870's, going past 900 core can be a pain.

 

[tviceman]

I just gave the card a 20 minute run in afterburner with the memory up to 3600mhz, and all was solid. 76 degrees fahrenheit inside my home, and the card topped out at 79 degrees celsius with the fan at 60%. I'll keep an eye on out for artifacts in games but, assuming this overclock holds, I'm very happy with these results and will keep it at this.