Gtx 780ti to be faster than Titan, will battle R9 290x

[boxleitnerb]

Then go ahead and test it with different CPU clocks. It won't change by more than 1-2%.

[tajoh111]

thats not 100% true in my opinion

IT shouldn't change much at all, particularly the extreme setting. Extreme setting basically ignore CPU power for the GPU part of the score. If this was performance setting I would agree it has an affect.

This is the newest 3dmark, but it basically shows this. BUt 3dmark extreme setting has always typically been good at getting rid of the CPU factor with earlier editions too.

http://www.ocaholic.ch/modules/smart...mid=948&page=4

Scores changed less than 1 percent from enabling 1 core to 4 cores on a 3770k.

[Rob94hawk]

There is no reason to release a 780Ti besides for epenis effect. Sure the 290X is good but it is hotter and lounder than Titan, therefore the latter is still more competitive and going by Nvidia's perceived image, could still sell for $100 more than the 290X. To me this is a kneejerk reaction to add a few frames more than AMD and keep the "fastest card" in the market. It is also a clear sign that the mystical Maxwell is no where to be found any soon, and most likely will show up only on 20nm in late 2014....

This is why I only build new rigs once every 5 to 6 years. When that time comes around the new hardware is 5 to 10 times faster. Not in increments of small percentages for massive amounts of cash.

[PedantOne]

thats not 100% true in my opinion

score of 290x is not important here, important is difference between Titan and 780 Ti score on the same system.

[behrouz]

Don't Worry ,Because R9 290 is around lower then 1% of the whole Products , so this won't hurt NV and Nv is not Worried for AMD R9 290x's Price.they have enough money to cover their loss.

[raghu78]

the 3dmark11 scores for gtx 780 ti seem to be 5% more than gtx titan. the problem is the R9 290X (uber) is on avg anywhere from 3 - 12% faster than GTX Titan at 1440p / 1600p depending on game suite, settings etc

http://www.hardwarecanucks.com/forum...review-18.html
http://www.techpowerup.com/reviews/AMD/R9_290X/27.html
http://techreport.com/review/25509/a...rd-reviewed/13
http://hexus.net/tech/reviews/graphi...9-290x/?page=5
http://www.gamestar.de/hardware/graf...3029260,5.html
http://www.purepc.pl/karty_graficzne...ieznanej_firmy
http://www.hardware.fr/articles/910-...formances.html

most likely Nvidia will share the GPU crown for the first time even with their big die GPU.

[boxleitnerb]

But only because Titans temperature target is set at 80C. If you compare quiet vs stock mode (and properly preheat) or uber mode vs power/temperature targets@max, they perform equally:
https://www.computerbase.de/artikel/...90x-im-test/5/
http://ht4u.net/reviews/2013/amd_rad...ew/index46.php

[Hornet331]

most likely Nvidia will share the GPU crown for the first time even with their big die GPU.

you seem to forget the 9700pro. who beat the FX series until the refresh of it.

[trinibwoy]

It stinks a bit when NVidia know they are going to release a secret revision (780Ti) that makes all other cards they happily took a lot of money for worthless. Thanks.

When a new card gets released all the cards before it don't magically get slower. Of course, if the only reason you buy graphics cards is for the ego boost of "having the fastest!" then yeah it's worthless.

[raghu78]

But only because Titans temperature target is set at 80C. If you compare quiet vs stock mode (and properly preheat) or uber mode vs power/temperature targets@max, they perform equally:
https://www.computerbase.de/artikel/...90x-im-test/5/
http://ht4u.net/reviews/2013/amd_rad...ew/index46.php

the way I see it GTX 780 Ti is going to trade blows with R9 290X . it comes down to what the partners are going to do with both these cards and how high the cards can go on custom cooler designs on air (toxic,lightning,matrix) and how high they clock on water (for watercooling enthusiasts). these are the only solutions which can be run 24/7.

[PatRaceTin]

VGA Card Naming is confuse me again

[VoodooProphetII]

Umm... both the Gtx 780 and Titan in those benchmarks are running 3930k and quad channel memory..... the R9-290x is running a 3770k....
Find some benchmarks running on the same system.

6 core 12 threats with 119xx Physics score compared to 125xx physics score with a 3770K 4 core 8 threats?

[boxleitnerb]

the way I see it GTX 780 Ti is going to trade blows with R9 290X . it comes down to what the partners are going to do with both these cards and how high the cards can go on custom cooler designs on air (toxic,lightning,matrix) and how high they clock on water (for watercooling enthusiasts). these are the only solutions which can be run 24/7.

As I said, the Titan and 290X already trade blows. The 780 Ti will be a tad faster, but possibly only equal clock for clock since it seems to be "just" a Titan with 2688 cores.

[PedantOne]

780 Ti will be faster then 290X in most of tests and games, 6GB VRAM will help in many HD+ resolutions. Be sure of this, Nvidia will compare 290X performance in many games and after that they will make final clock decision. But with lower power consumption, less noise, lower temps a with 10-15 percents OC headroom.

780 Ti will be winner in ALL aspects, except the price i think.

[raghu78]

780 Ti will be faster then 290X in most of tests and games, 6GB VRAM will help in many HD+ resolutions. Be sure of this, Nvidia will compare 290X performance in many games and after that they will make final clock decision. But with lower power consumption, less noise, lower temps a with 10-15 percents OC headroom.

780 Ti will be winner in ALL aspects, except the price i think.

at 4K Ultra HD or multi monitor the gap is significant. that 512 bit memory bus and 64 ROPs really make a difference at ultra high resolutions. clock speeds alone cannot make up for architectural differences.

http://www.hardocp.com/article/2013/...card_review/14
http://www.nordichardware.se/Grafikk...0.html#content
http://www.tomshardware.com/reviews/...ew,3650-9.html
http://www.anandtech.com/show/7457/t...290x-review/11

[boxleitnerb]

at 4K Ultra HD or multi monitor the gap is significant. that 512 bit memory bus and 64 ROPs really make a difference at ultra high resolutions. clock speeds alone cannot make up for architectural differences.

http://www.hardocp.com/article/2013/...card_review/14
http://www.nordichardware.se/Grafikk...0.html#content
http://www.tomshardware.com/reviews/...ew,3650-9.html
http://www.anandtech.com/show/7457/t...290x-review/11

No it is not. You're still forgetting that the GeForce cards don't run with maximum clocks, but the Radeons do, often in quiet mode as well (as evidenced by the negligible difference between quiet and uber mode in many cases, showing insufficient pre-heating procedures). This is due to the 80C temp target, especially now when many sites have changed their testing procedure to warm up the cards. I have provided you with two reviews that do proper comparisons with Titans "uber mode" if you will. The difference in 4K on average is minimal.

It's funny how incompetent many reviewers are. First they benchmark Nvidia boost 2.0 cards without pre-heating, creating unrealistic testing conditions. Now that AMD has implemented a temperature-dependent boost, too, reviewers test AMDs quiet AND uber mode but only test Nvidias "quiet" (stock) mode, yet again creating an unfair comparison.

[tbone8ty]

so if Amd released a slower R9-290x but more power efficient, quieter, and lower temps than a 780. people would still complain

[boxleitnerb]

The only complaint is about untrue statements. It's funny since raghu78 always wants so see clock for clock results, but when they disprove what he says, he ignores them

[LordEC911]

so if Amd released a faster R9-290x but more power efficient, quieter, and lower temps than a 780. people would still complain

Fixed it for you.
The people complaining don't care about the product, they care about their agenda.

FYI- Hawaii DP is 1/2... it is Radeon neutered by software.

[raghu78]

No it is not. You're still forgetting that the GeForce cards don't run with maximum clocks, but the Radeons do, often in quiet mode as well (as evidenced by the negligible difference between quiet and uber mode in many cases, showing insufficient pre-heating procedures). This is due to the 80C temp target, especially now when many sites have changed their testing procedure to warm up the cards. I have provided you with two reviews that do proper comparisons with Titans "uber mode" if you will. The difference in 4K on average is minimal.

what. the majority of reviews show performance difference on avg of 10 - 20% at 4k. so the difference is not minimal. btw there are reviews which found even uber mode throttling clocks in few games. so again you are wrong.

http://www.hardocp.com/article/2013/...3#.Um1CU1P098E

It's funny how incompetent many reviewers are. First they benchmark Nvidia boost 2.0 cards without pre-heating, creating unrealistic testing conditions. Now that AMD has implemented a temperature-dependent boost, too, reviewers test AMDs quiet AND uber mode but only test Nvidias "quiet" (stock) mode, yet again creating an unfair comparison.

btw nvidia does not provide a uber mode. you have to manually work the fan settings. so who determines what fan settings to run ? instead of accusing reviewers read the reviews properly. study their benchmarking procedures and then take the results from the ones you consider fair.

http://www.hardwarecanucks.com/forum...-review-5.html

"Now, before every run, cards are left to heat up for 10 minutes under in-game load which will ensure more realistic benchmark numbers. "

The only complaint is about untrue statements. It's funny since raghu78 always wants so see clock for clock results, but when they disprove what he says, he ignores them

yeah i am all for clock for clock comparisons which seem impossible unless both Nvidia and AMD allow you to disable boost completely and just work with clocks, voltages and fan speeds with a hard temp limit (throttle if exceeded) for safety reasons.

[AliG]

780 Ti will be faster then 290X in most of tests and games, 6GB VRAM will help in many HD+ resolutions. Be sure of this, Nvidia will compare 290X performance in many games and after that they will make final clock decision. But with lower power consumption, less noise, lower temps a with 10-15 percents OC headroom.

780 Ti will be winner in ALL aspects, except the price i think.

Unless if they give a 512 bit memory controller I'm not sure they can win out at the higher resolutions. After getting a 1440p monitor there's no way I'm ever going back to 1080p, and I think a lot of others feel the same now that they are becoming more affordable

[STEvil]

only way i'd go back to 1080p is if it were 2560 pixels wide

[vario]

only way i'd go back to 1080p is if it were 2560 pixels wide

There are monitors like that ;-)

All this shows just how much some competition can get us, man it would be great if AMD could get their act straight on the cpu front.

780ti will with all possibility be faster on average than R290x, thats a given, they have the silicon, and they have the headroom.
But after 290x, they cant really go overboard with price, they either attack AMD at their price point, which will mean AMD will lower theirs, or go for the 780 price point, they cant really get higher.
Titan is obsolete, of course, they can do a titan ti ,and keep the price points.All this makes it pretty possible for me to buy hawaii LE in few months

[tajoh111]

Fixed it for you.
The people complaining don't care about the product, they care about their agenda.

FYI- Hawaii DP is 1/2... it is Radeon neutered by software.

Really, that's just troll bait. Everyone would be happy if it was faster, cheaper, quieter and cooler. That would just be a flawless product.

Consuming more power and creating more heat in the process than a gtx 480(techpowerup and guru3d reviews, even in quiet mode) is not a good thing and likely makes a new record for a single GPU card. Every one, particularly AMD fans complained about the heat problems of the gtx 480.

So when a 290x, I would expect the same criticism. However, with the 290x, it doesn't even fully capture the performance crown like the gtx 480 but to its benefit, it is priced much better than its competition.

For a company that has championed efficiency for a while now, I would hope AMD could create a more efficient chip after having more than two years to work on this process(and during this course they made a whole line). With GF100, there was a good chance that such a monster die as their first product on 40nm was going to be a leaky processor. So AMD really has no excuse for the problems of Hawaii. Well Nvidia should have made such a big die as their first chip either so it is their fault too. But with AMD making quite a bit smaller product, and having that experience, they should have made something a bit more efficient.

Its seems like a problem with GCN in general and left them with two choices. Increase the utilization of the GCN cores and power goes up a lot or just make a monster die like Nvidia. They went with sort of both. But it doesn't have the max performance of a monster die but has the power usage of one. They got part of the performance but the entire cost of the power consumption.

Fermi was already breaking or at the limit of how much energy and how hot a processor can be. To go 10-15 percent beyond this isn't the direction AMD can continue with. Nvidia made huge improvements with keplar and likely changed their design scheme to not repeat past mistakes of Fermi. I.e making gx104 first then the big die and just generally making a more efficient architecture.

I find it really strange that people think the practical solution to this bad cooler(which is universally being called crap) debacle is to buy a waterblock and thus a water loop. The fact that the best solution is water cooling which is kind of extreme and atleast 200 dollars in water cooling(125 dollar block, radiators, tubes and all that stuff or a cheap water cooling kit + block) kind of shows how much people are willing to overlook AMD. The extra cost really kills the savings and the default cooler should be better even if it meant raising the cost by 20 bucks.

As a result of this bad cooling solution, CF crippled by air as open air cooling + high heat producing product makes water almost essential. If AMD had put a better blower fan on it, it would have atleast killed the noise issue and made crossfire more usable. As is, it seems like AMD is forcing people to make too many compromise unless they are willing to spend huge amounts on cooling.


Yet it's tough to recommend R9 290X CrossFire because it's anything but quiet. No matter whether you use the "Quiet" or "Uber" BIOS, they're both unbearably loud when properly loaded and not recommendable for setups in which you rely on speakers, no matter how punchy they are. Their noise output can even potentially affect your gameplay in online multi-player games where you can hear out your enemies. We'd strictly recommend a gaming headset with good noise isolation, preferably active noise cancellation.

http://www.techpowerup.com/reviews/A...ssFire/23.html

The point where loud speakers are not enough and preferably noise canceling headphone is ridiculous.

[AliG]

only way i'd go back to 1080p is if it were 2560 pixels wide

I dunno dude, I looked at some 21:9 monitors, and it's frankly an awkward experience. I think 1440p is really the way to go, especially with the QX2710 having an overclockable version of the same Samsung Matte PLS panel in many high end professional monitors.