Synthetics

As always we’ll also take a quick look at synthetic performance to see if NVIDIA’s core configuration has had any impact on basis performance metrics. We’re expecting to see performance very close to the GTX 660 Ti, due to its nearly identical ROP/L2/memory configuration. We’ll start with 3DMark Vantage’s Pixel Fill test.

3DMark Vantage’s pixel fill test likes memory bandwidth and ROP performance in that order, which makes these results a bit odd. With identical memory bandwidth between them we’d expect the GTX 660 and GTX 660 Ti to at least be tied here, if not a slight lead for the GTX 660 thanks to its higher ROP performance. Instead the GTX 660 trails the GTX 660 Ti by a slight amount, an outcome we can’t explain at this time.

Our texture fillrate benchmark on the other hand sees a large gap between the GTX 660 and GTX 660 Ti, which is what we would expect from the loss of SMXes.

Our third theoretical test is the set of settings we use with Microsoft’s Detail Tessellation sample program out of the DX11 SDK

Despite the loss of SMXes (and thereby Polymorph engines), our tessellation benchmarks don’t show any kind of significant difference between the GTX 660 and GTX 660 Ti. We’ve been finding this benchmark to be surprisingly sensitive to ROP performance and memory bandwidth on Kepler, and these results back that finding.

Our final theoretical test is Unigine Heaven 2.5, a benchmark that straddles the line between a synthetic benchmark and a real-world benchmark as the engine is licensed but no notable DX11 games have been produced using it yet.

Despite its advanced nature, Heaven isn’t particularly sensitive to the loss of shader and texturing performance, as signified by the performance loss of less than 10% for the GTX 660.

Compute Performance Power, Temperature, & Noise
Comments Locked

147 Comments

View All Comments

  • TemjinGold - Thursday, September 13, 2012 - link

    "For today’s launch we were able to get a reference clocked card, but in order to do so we had to agree not to show the card or name the partner who supplied the card."

    "Breaking open a GTX 660 (specifically, our EVGA 660 SC using the NV reference PCB),"

    So... didn't you just break your promise as soon as you made it AND show a pic of the card right underneath?
  • Sufo - Thursday, September 13, 2012 - link

    Haha, shhhh!
  • Homeles - Thursday, September 13, 2012 - link

    Reading comprehension is such an endangered resource...

    If it's the super clocked edition, it's obviously not a reference clocked card.
  • jonup - Thursday, September 13, 2012 - link

    Exactly my thoughts.
  • Ryan Smith - Thursday, September 13, 2012 - link

    Homeles is correct. That's one of the cards from the launch roundup we're publishing later today.. The reference-clocked GTX 660 we tested is not in any way pictured (I'm not quite that daft).
  • knutjb - Saturday, September 15, 2012 - link

    No matter what you try to say it still reads poorly. It should be blatantly obvious about which card was which up front, which the article wasn't. I should have to dig when scanning through.

    Also, your picking it as the better choice over a card that has been out how long, over slight differences... If nvivda really wanted to me to say wow I'll buy it now, the card would have been no more than 199 at launch. 10 bucks under is the best they can do for being late to the party? And you bought the strategy. I have been equally disappointed with AMD when they have done the same thing.
  • MrSpadge - Sunday, September 16, 2012 - link

    When reading Anadtech articles it's almost always safe to assume "he actually means what he's saying". Helps a lot with understanding.
  • thomp237 - Sunday, September 23, 2012 - link

    So where is this roundup? We are now 10 days on from your comment and still no signs of a roundup.
  • CeriseCogburn - Friday, October 12, 2012 - link

    I have been wondering where all the eyefinity amd fragglers have gone to, and now I know what has occurred.

    Eyefinity is Dead.

    These Kepler GPU's from nVidia all can do 4 monitors out of the box. Sure you might find a cheap version with 3 ports, whatever - that's the minority.

    So all the amd fanboys have shut their fat traps about eyefinity, since nVidia surpassed them with A+ 4 easy monitors out of the box on all the Kelpers.

    Thank you nVidia dearly for shutting the idiot pieholes of the amd fanboys.

    It took me this long to comment on this matter because nVidia fanboys don't all go yelling in unison sheep fashion about stuff like the little angry losing amd fans do.

    I have also noticed all the reviewers who are so used to being amd fan rave boys themselves almost never bring up multimonitor and abhor pointing out nVidia does 4 while amd only does 3 except in very expensive special cases.

    Yeah that's notable too. As soon as amd got utterly and totally crushed, it was no longer a central topic and central theme for all the review sites like this place.

    That 2 week Island vacation every year amd puts hundreds of these reporters on must be absolutely wonderful.
    I do hope they are treated very well and have a great time.
  • EchoOne - Wednesday, November 21, 2012 - link

    LOL dude,the 660ti vs the 7950 in eyefinity would get destroyed.I know this because my friend has a comp build with a phenom 965be 4.2ghz and 660ti with 16gb of ram (i built this for him) and i have a fx 6100 4.7ghz,16gb ram and a 7950 i run a triple monitor setup

    https://www.youtube.com/watch?v=ZRXGveviruw&fe...

    And his 660ti DIED trying to play the games at that res and at the same settings as i do.He had to take down his graphics settings from say gta4 from max settings down to about medium and high (i run very high)

    So yeah sure it can run a couple monitors out of the box but same with eyefinity.And trust me their nvidia surround is not as polished as eyefinity..But they get props for trying.

Log in

Don't have an account? Sign up now