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

[24601]

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

Lower stock voltages = lower overclocks.

Duh.

When cheating a power consumption and throttling test you need to give something up.

Also, well played on burying my information post to the next page.

Now I just have to wait for the deluge of viral marketeers to astroturf my post beneath 2 pages of spam again. Quite fun.


Personal attacks and use of inflammatory language is against the forum rules. You can take the next 3 days off for this.

-Rvenger

 

[BallaTheFeared]

Lower stock voltages = lower overclocks.

Duh.

When cheating a power consumption and throttling test you need to give something up.

Also, well played on burying my information post to the next page.

You also forgot the part where users lie, users do 1 minute benchmarks, users use 100% fan speed, users modify bios for unsupported voltage.

This one?

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.

 

[24601]

You also forgot the part where users lie, users do 1 minute benchmarks, users use 100% fan speed, users modify bios for unsupported voltage.

Yes, its funny that OCN claims 1300 is "easy" to reach on 290x.

 

[BallaTheFeared]

It's always easier if your card can actually do it.

 

[blackened23]

Yes, its funny that OCN claims 1300 is "easy" to reach on 290x.

The only thing OCN users care about is doing suicide valley and 3d mark runs. None of their results have any basis in reality, because literally 95% of those guys do suicide runs that aren't stable and have artifacts. Yet if they get a score, that score counts.

They might get 1300mhz for like. Two minutes. They may even finish a 3dmark run. BIG MAYBE.

But it sure as heck isn't going to be 24/7 stable. Nobody at OCN cares about 24/7 stable overclocks. Just suicide runs all day long, therefore all of those results are worthless for judging real world performance.

 

[RaulF]

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.

what does that have to do with holding clock? And the game runs is stable? Yes via ccc i raised max fan speed. Also lowered temp to. Everyone can do that.

Your point is?

 

[BallaTheFeared]

Maybe nothing, like your post as it pertained to the issues discussed in the OP.

Or maybe something, who knows?

 

[chimaxi83]

The only thing OCN users care about is doing suicide valley and 3d mark runs. None of their results have any basis in reality, because literally 95% of those guys do suicide runs that aren't stable and have artifacts. Yet if they get a score, that score counts.

They might get 1300mhz for like. Two minutes. They may even finish a 3dmark run. BIG MAYBE.

But it sure as heck isn't going to be 24/7 stable. Nobody at OCN cares about 24/7 stable overclocks. Just suicide runs all day long, therefore all of those results are worthless for judging real world performance.

Proof? Source? Anything?

I'll wait.

 

[Erenhardt]

The only thing OCN users care about is doing suicide valley and 3d mark runs. None of their results have any basis in reality, because literally 95% of those guys do suicide runs that aren't stable and have artifacts. Yet if they get a score, that score counts.

They might get 1300mhz for like. Two minutes. They may even finish a 3dmark run. BIG MAYBE.

But it sure as heck isn't going to be 24/7 stable. Nobody at OCN cares about 24/7 stable overclocks. Just suicide runs all day long, therefore all of those results are worthless for judging real world performance.

Reviewers can't be bothered to test second bios setting. Sometimes they don't test overclocking at all. Is there any reason we should think they make sure their overclocks are 24/7 stable?

 

[RaulF]

Maybe nothing, like your post as it pertained to the issues discussed in the OP.

Or maybe something, who knows?

The issue is Are press samples faster than retail!

I have a retail card. It holds clock.

And just so you and other AMD haters can't say i modified how the card works from the factory.

Heres a screen shot i just did again, BF4 21 minutes total game play in Shangai.

I raised target temp back to 95, returned fan to 47 max.

So i returned it on how it was design to work from AMD direct.



And before you bring up the gpu is going crazy, yeah if it has nothing to do it will try to down clock. Also this is multi player, i guess when [redacted] goes crazy explosions and what not it will use more gpu.

Warning issued for inappropriate language.
-- stahlhart

 

[ICDP]

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.

Explain why Asus ROG and MSI Lightning 7970s were almost always in 60s for ASIC? Lower ASIC has almost always allowed higher OC potential because they can handle more volts without crapping out.

 

[BallaTheFeared]

Still over 47% :|

It's my AMD hater ways that compel me to notice these things.


Given the need to go over the 47% cap, the erratic usage, and the possibility of open windows I'll assume based on AMD hate that your card throttles.

 

[Final8ty]

Explain why Asus ROG and MSI Lightning 7970s were almost always in 60s for ASIC? Lower ASIC has almost always allowed higher OC potential because they can handle more volts without crapping out.

My ASIC 73.6% runs hotter than my ASIC 56.0% and ASIC 60.4%. 7950's

 

[RaulF]

Still over 47% :|

It's my AMD hater ways that compel me to notice these things.


Given the need to go over the 47% cap, the erratic usage, and the possibility of open windows I'll assume based on AMD hate that your card throttles.

Show me the throttle?

I see a 94X clock speed steady.

 

[ICDP]

My ASIC 73.6% runs hotter than my ASIC 56.0% and ASIC 60.4%. 7950's

My 70% ASIC 7950 could do 1230 core while my 90% ASIC one could do 1170 at best. Even then putting any reasonable amount of volts through it made it overheat like crazy. I had an MSI lightning GTX680 with an ASIC in mid 60s. Yet now we are to believe that ASIC is the sole determinant when it comes to quality and OC potential.

Asus, MSI, Gigabyte etc must be crazy to bin lower ASIC chips for their elite cards. :whiste:

 

[SiliconWars]

ASIC "quality" is complete BS.

http://forum.beyond3d.com/showpost.php?p=1808073&postcount=2130

 

[blastingcap]

I used to believe in ASIC quality, but not anymore. I've seen twice now where the higher-ASIC quailty chip allowed for slightly lower voltage in Afterburner/TRIXX, but when I measured the wattage with a Killawatt, I was shocked to see that the lower-voltage card ate more power. I tried again with another pair of cards and same thing. Granted, these were cards from different product lines, so it is possible that higher ASIC quality means something within the SAME product line, so maybe a 80% ASUS DCuII takes less wattage than a 70% one, for instance. But a 80% ASUS DCuII 7790 may actually eat more wattage than a 60% 7790 MSI one. This is the beginning of my hatred of smaller fans, by the way, because in both cases I've measured, the smaller-fanned card ate more wattage than the bigger-fanned card despite the bigger-fan card having higher voltage, so maybe cooling efficiency had something to do with it. If I had it my way, I would rather have one massive 120mm fan than two 92mm fans.

 

[blackened23]

You know, I would have to agree with the ASIC score not having a real correlation to chip quality. I have yet to find a direct correlation between a high ASIC and good overclocking with the cards i've owned, just my personal experience. I've had low cards OC high and vice versa.

 

[3DVagabond]

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.

I suppose you call this some kind of proof that AMD is cheating? Let's assume this post is true and not something someone made up (Which we have zero proof of the credibility of the info you've posted here.

1, Show me an informed review based on asic quality that demonstrates any correlation between asic quality % and any performance measurement.

2, Show me any legit study that shows the asic for consumer cards is lower relative to review cards.

3, Clicking your link I don't find the pick of the press sample you show. I've checked a couple of posts that it's linked to and can't find it either. I assume you are saying the pick that shows the GPUz of the review sample you've pictured. Can you link to show that?

4, Are you accusing AMD of shenanigans, or the AIB partners of using substandard chips in their premium cards? Your last sentence seems to be insinuating the latter, even though we haven't even had a custom card released yet.

 

[zhuks]

Here is my Sapphire 290x BF4 edition on water. Running 1150mhz core and 1450mhz memory. Stock voltage. Played 2 full conquest maps in BF4 and you can see there is 0 throttle. From what I can tell in terms of GPU usage it only drops when you die ect.

I normally run higher with additional voltage, but wanted to show even an overclocked card properly cooled with no additional voltage can run at a good amount above stock with 0 throttle.

 

[Face2Face]

^ How high have you got your card on water with extra voltage?

 

[zhuks]

^ How high have you got your card on water with extra voltage?

1230mhz core and 1570mhz on memory with +100mv
1290mhz core and 1570mhz on memory with +138mv

I need more voltage for 1300mhz+, but I also need to use something other than MSI afterburner since my slider maxes out at 1300mhz. As for memory I can max out the slider as well at 1625mhz, but I have seen no performance increase above 1500-1550mhz so I keep it at 1570mhz.

And for those interested 75.6% ASIC.