Are retail 290x cards slower than press samples? (Tech Report)

[SlickR12345]

Useless review. First they run the cards in quiet mode which will throttle the cards, second its obvious the HIS card for some reason the cooler runs at a significantly lower RPM and thus has lower frequency and lower performance.

Plus I've noticed that they run the tests outside a case and in normal room, which based on how the cards work is stupid.

You need to run them in a case with a stable temperature and have the room temperature properly controlled as well and always the same temperature so you can actually get proper results.

There is no doubt the press cards perform slightly better in quiet mode, insignificantly less so in uber mode.

Its clear that the big issue is the cooler. I mean if the cards has a proper cooler that was able to keep them cool and at low noise, all cards would perform significantly better.

In fact I think its worthwhile to buy an aftermarket cooler for $30 to $50 and replace the AMD cooler on these cards.

 

[parvadomus]

Yes they cant touch AMD shit performance first time in ur life u said a truth.

I dont understand ur logic.Bro if u wake up than u will see reality that Nvidia always won the gpu war and majority of user due to AMD lies and false hype.


Nvidia
GTX 780TI>R9 290 and R9 290X

Fact is AMD always copied Nvidia example are SLI,Shadow play etc.

Are you a nv fanboy? Follow the topic please. NV cards are garbage for mining as vagabond pointed, and that is a FACT.

Infraction issued for personal attack.
-- stahlhart

 

[KCfromNC]

just look at any Kepler specs on any website, rather that be amazon, newegg, techpowerup, anandtech or wherever. it will bluntly tell you the guaranteed base clock.

for example. 680 is guaranteed at 1006MHz. will oc to 1150MHz. if you really lucky with the silicon lottery, then to 1225MHz.

any gpu that does not meet that guarantee base clock simply gets a rma as defective.

any more questions?

Sure, the same one I asked in the post you responded to - where can I find reviews of the performance of NV cards running at their guaranteed base clocks. All I can find anywhere are reviews of them running at max boost clocks, just like all the reviews of AMD cards.

In other words, guaranteed base clocks are irrelevant if the performance of the cards is advertised at peak boost clocks. NV's written a number down somewhere on a data sheet or something, but that number is totally irrelevant to the conversation about performance.

 

[blackened23]

I would rather they didn't waste their limited resources catering to your every whim and just continue to bring the value ()🙂

It's not just my needs, I think a lot of users finds consistent performance (without variance, or press vs retail differences) with excellent user experience worthwhile. AMD has the ability to do this - if you recall the 5870, I considered that to be a better product than what nvidia offered at the time. It had a better user experience than the GTX 480 and it was a better value at the same time; user experience metrics (acoustics, software etc) and value aren't mutually exclusive - they can all co-exist. While the GTX 480 was slightly faster, it cost quite a bit more and the user experience was worse than what the 5870 offered. Basically, the 480 was an overly hot and loud card that a lot of people didn't care for. Right now, I see the tables as turned and the 290X is basically AMD's Fermi/GTX 480: I do feel that the 780oc offers a better overall experience than the 290 or 290X by a large margin. The performance may be similar but the other stuff? Software, acoustics, overall user experience favors the 780 by a lot. (my opinion, and yes, it's subjective.) I'd love it if AMD could course correct and make the GPU that is great by every metric, with consistent performance between press/retail.

 

[Firestorm007]

First off there are no numbers to back your statement that the largest contingent of AMD card buyers don't play games. Also all of this, it's not sustainable, nay saying has been going on since BTC first came along years ago, but mining for cryptocurrency is still going strong and is arguably stronger than ever. People who played their cards well have made a lot of money over this time. History denies your view on the whole subject. There's really no evidence that it's going to ever be unsustainable at this point in time.

Wow...Just wow...I agree 3D. I don't know where to begin on that one. I'm sorry, but that's just an idiotic statement with zero basis in fact. I'd love to know where that fairy tail came from. I know a lot of people who play on AMD cards, actually most of them do. So, just because a lot of people mine on AMD cards=that people don't play games as well. That's some serious logic fail or something else entirely. Nevertheless, I would like some proof of this statistic?

Warning issued for personal attack.
-- stahlhart

 

[OatisCampbell]

Wow...Just wow...I agree 3D. I don't know where to begin on that one. I'm sorry, but that's just an idiotic statement with zero basis in fact. I'd love to know where that fairy tail came from. I know a lot of people who play on AMD cards, actually most of them do. So, just because a lot of people mine on AMD cards=that people don't play games as well. That's some serious logic fail or something else entirely. Nevertheless, I would like some proof of this statistic?

About all I do with AMD cards is play games. I could use the integrated graphics on a cpu for everything else I do.

About the only thing the 290 is going to give me over my 7970GHz is more games with playable AA at 25x16.

 

[UaVaj]

Sure, the same one I asked in the post you responded to - where can I find reviews of the performance of NV cards running at their guaranteed base clocks. All I can find anywhere are reviews of them running at max boost clocks, just like all the reviews of AMD cards.

In other words, guaranteed base clocks are irrelevant if the performance of the cards is advertised at peak boost clocks. NV's written a number down somewhere on a data sheet or something, but that number is totally irrelevant to the conversation about performance.

🙄

 

[thedosbox]

Useless review. First they run the cards in quiet mode which will throttle the cards, second its obvious the HIS card for some reason the cooler runs at a significantly lower RPM and thus has lower frequency and lower performance.

Which is the stock experience a retail customer would get. For now at least.

Plus I've noticed that they run the tests outside a case and in normal room, which based on how the cards work is stupid.

Running the cards outside of a case should help temperatures no?

Its clear that the big issue is the cooler. I mean if the cards has a proper cooler that was able to keep them cool and at low noise, all cards would perform significantly better.

In fact I think its worthwhile to buy an aftermarket cooler for $30 to $50 and replace the AMD cooler on these cards.

Hence the recommendation to wait for the custom coolers. Welcome to page 1 of the thread.

Having said that, you're suggesting someone void their warranty at an additional cost of $50? Really?

 

[thilanliyan]

Having said that, you're suggesting someone void their warranty at an additional cost of $50? Really?

Not all companies void warranties for use of an aftermarket cooler. XFX doesn't for example and I know there are a couple of others on the AMD side.

 

[desprado]

Are you a nv fanboy? Follow the topic please. NV cards are garbage for mining as vagabond pointed, and that is a FACT.

Infraction issued for personal attack.
-- stahlhart

U have to learn first what the topic is saying and if u cannot read than than google it translate in own language.

 

[desprado]

Sorry for posting of the topic.
Moderators and sorry for personal attack.

 

[Will Robinson]

This whole issue of manipulating the Kepler GPUs fan setting "just because " is utterly absurd. Fact of the matter is, Kepler GPUs meet 99-100% of their performance potential @ factory default auto fan settings without ever touching the fan speed. The throttling is little to none with 1-2 bin throttling of 13-26mhz. The only reason to touch fan settings is for better acoustics or better performance. Even at auto fan settings, Kepler is not having drastic performance variances.

Wasn't there a quote and chart from one of the other big websites saying that during testing the 78oTi had to be run at 80% fan speed to keep it at 1000mhz?

 

[KCfromNC]

🙄

Here's what happens when marketing talking points run headfirst into reality.

 

[thedosbox]

Not all companies void warranties for use of an aftermarket cooler. XFX doesn't for example and I know there are a couple of others on the AMD side.

Thanks, I did not know that. That's impressive customer relations from XFX.

Still, having to spend an extra $30 - 50 would reduce the value proposition for me. But then, I've already upgraded from a 7870 so it's of academic interest anyway.

 

[UaVaj]

Here's what happens when marketing talking points run headfirst into reality.

ignorance is bliss.

🙄🙄

 

[RaulF]

Here to fire the flames.

This is my ref sapphire 290. Notice the clock is steady, now look at temp and far speed.

So this whole the cards are downclocking [redacted] is BS. At least on my side. I order another one yesterday and once it gets here i will post results of it too.

Profanity isn't allowed in the technical forums.
-- stahlhart

 

[ICDP]

Here to fire the flames.

This is my ref sapphire 290. Notice the clock is steady, now look at temp and far speed.

So this whole the cards are downclocking [redacted] is BS. At least on my side. I order another one yesterday and once it gets here i will post results of it too.

R9 290s are not affected, only R9 290X and only in quiet mode. Essentially 40% fan speed is not enough to maintain 1000MHz and depending on the card it can throttle by quite a margin. It should be remembered that some cards with lower ASIC can use more volts and as such can throttle more as they heat up faster.

It has been shown that in Uber mode the variance is within normal testing variance.

Don't want throttling switch to Uber mode.

http://techreport.com/blog/25751/an-update-on-radeon-r9-290x-variance

 

[RaulF]

R9 290s are not affected, only R9 290X and only in quiet mode. Essentially 40% fan speed is not enough to maintain 1000MHz and depending on the card it can throttle by quite a margin. It should be remembered that some cards with lower ASIC can use more volts and as such can throttle more as they heat up faster.

It has been shown that in Uber mode the variance is within normal testing variance.

Don't want throttling switch to Uber mode.

http://techreport.com/blog/25751/an-update-on-radeon-r9-290x-variance

Seriously!!!!

I haven't read every post here or the fine details of the "article". They are bring quiet mode and talking about throttling. 🙄

Some delusional individuals testing cards now days.

 

[ICDP]

Seriously!!!!

I haven't read every post here or the fine details of the "article". They are bring quiet mode and talking about throttling. 🙄

Some delusional individuals testing cards now days.

It should also be noted that Nvidia bought them the retail R9 290X cards to test.

admittedly R9 290X in quiet mode does throttle depending on thermal levels but this is by design and will vary from card to card. It is bad that they do this and IMHO AMD deserve all the crap they get for the inadequate reference cooler. Having said that these articles are pure Nvidia marketing.

 

[Face2Face]

R9 290s are not affected, only R9 290X and only in quiet mode. Essentially 40% fan speed is not enough to maintain 1000MHz and depending on the card it can throttle by quite a margin. It should be remembered that some cards with lower ASIC can use more volts and as such can throttle more as they heat up faster.

I have noticed with my high Asic 7950's that they actually draw more wattage (when overclocked) and run hotter from the get go vs. Lower ASIC 7950 cards. Could be because of the reference 7970 board or due to being a less leaky chip?

 

[krumme]

It should also be noted that Nvidia bought them the retail R9 290X cards to test.

admittedly R9 290X in quiet mode does throttle depending on thermal levels but this is by design and will vary from card to card. It is bad that they do this and IMHO AMD deserve all the crap they get for the inadequate reference cooler. Having said that these articles are pure Nvidia marketing.

Yes its a farce. There is nothing new in this thread that was not known before. Yet we have to read all this misleading stuff paid by and promoted by NV. And when this thread is critizided we are told to stay on topic.

What topic? The one set by nv marketing? Catch 22.

The thread title is misleading. The article was bended and misleading from day one. There have been plenty argument for it. Time to close this pr machine.

 

[BallaTheFeared]

Here to fire the flames.

Your usage looks like your video card went full retard.


Also you're going over 47% so you've modified how the card works anyways.

 

[Face2Face]

I thought there was a bug in MSI Afterburner that doesn't show GPU usage for R290/X properly?

 

[24601]

Press samples have much better binned chips.

AMD does this all the time

http://www.overclock.net/t/1436497/official-amd-r9-290x-290-owners-club/6620

End-user's ASIC quality almost never goes over 80%, and most hover between 65-73%

Higher ASIC quality + lower voltage press BIOS

All bout paid reviewer marketers.

All the suckers buying AIB AMD/ATi cards should prepare for the failure-binned 50-60% ASIC quality chips on UBER PREMIUM ULTRA CUSTOM MILITARY CLASS boards.

 

[Erenhardt]

Press samples have much better binned chips.

AMD does this all the time

http://www.overclock.net/t/1436497/official-amd-r9-290x-290-owners-club/6620

End-user's ASIC quality almost never goes over 80%, and most hover between 65-73%

Higher ASIC quality + lower voltage press BIOS

All bout paid reviewer marketers.

All the suckers buying AIB AMD/ATi cards should prepare for the failure-binned 50-60% ASIC quality chips on UBER PREMIUM ULTRA CUSTOM MILITARY CLASS boards.

How that explains users reporting 100% better overclocks than what was achieved by reviewers?