[Hardware.fr] Asus, MSI and the "special press" bios

[dacostafilipe]

Original Source: http://www.hardware.fr/news/14693/gtx-10x0-asus-msi-bios-special-presse.html
Videocardz Source : http://videocardz.com/61121/asus-and-msi-accused-of-sending-modified-cards-to-the-press

I quote Videocardz as they already translated the important parts:

According Damien Triolet, a known and respected GPU reviewer (who, as you might remember, helped to unravel the secrets in GTX 970 memory allocation problem), ASUS and MSI are sending cards with modified BIOSes to the press. Such software enables more power on review samples, which leads to increased frequencies and better results overall.

Damien reports that those ‘optimized’ BIOSes are a very common problem in GPU industry. Manufacturers often encourage GPU reviewers to enable special overclocking presets before attempting to review those cards. Luckily, with little success.

The alternative is to supply optimized BIOSes to the press, so such settings are enabled by default. This usually means a gain of few MHz. Something that you won’t get on retail sample.

For such reason Damien asks manufacturers to supply retail BIOSes for his tests. Obviously manufactures are not eager to supply such software. The problem was discovered with MSI GTX 1080 GAMING X and ASUS GTX 1070 STRIX, so might want to take reviews of those cards with a grain of salt. Gigabyte on the other hand does not use such practices with its G1 GAMING Series.

Damien Triolet, Hardware.fr:

Obviously, Asus and MSI have made the calculation that getting these small gains brought more benefits than criticism among some troublemakers. Sure, after all, so these bios boost perhaps the performance of the cards by 1%, but that’s no reason to look elsewhere! After 1%, it will be what? 2%? Then 3%? Then widespread cheating contest?

 

[casiofx]

Moral of the story, buy the Gigabyte, vote with your money?

 

[lixlax]

I think its not a big surprise, just someone dug deeper into it. On the hardware level these samples probably have the best crerry picked chips as well for few additional MHz and/or less watts.

 

[kraatus77]

press only driver, press only bios. sooon only press will get a full product and retail will get something different. lol

we need better journalists. current once are just pathetic.

 

[LTC8K6]

Well, we know how to check and change video card BIOSes, don't we?

Also, this is why you should buy the card yourself to review it properly.

Cards given to you, or anything given to you, for testing and review, should be suspect.

 

[JimKiler]

I remember over 10 years ago it was discovered Asus did the same thing with their motherboards. They were a couple of MHz overclocking the CPU so their mobo was always the best in benchmarks.

 

[Silverforce11]

I remember over 10 years ago it was discovered Asus did the same thing with their motherboards. They were a couple of MHz overclocking the CPU so their mobo was always the best in benchmarks.

That's fine and good for custom cards... but not cool at all if it's actually a special mod bios for review samples only. That's cheating.

 

[Borealis7]

the OC profiles, at least in MSI's case, are available if you install their app.

TechPowerup are on it as well:
http://www.techpowerup.com/223440/m...-samples-with-higher-clocks-than-retail-cards

 

[zentan]

Asus has resorted to such practices even for their motherboards,nothing new.
Raise base frequency by a little just to be on top of table or MCE cheats,they have done all. And one of those brands which cheaps out in mid-tier and low ends line-ups at times. r9 290x DCII is just another example(a design for larger gk110 chips used for smaller Hawaii) of selling on the basis of name.
I am sure they would have sold more crappy 290 series DCII than Sapphire would have sold 290 series Tri-X.

 

[Tumaras]

The only way really around this is for reviewers to get their samples through retail channels. Otherwise since the beginning of time companies have been playing these games with cpus and video cards, whether it's golden samples, modified bios, modified driver, etc. It's so bad that I always knock a fair % off a review top oc number, since I know retail versions will always be a bit lower. This is just a rare case of it being noticed and called out.

 

[Silverforce11]

MSI...

 

[imaheadcase]

I like how his name is repeated constantly, as to like say "I'm so clever everyone look at me". Which in truth none of it really matters

The same with the 970 memory, it was a non issue.

 

[Face2Face]

MSI...

Pretty shady, but not unexpected.

I hope people are not buying MSI over Gigabyte or any other AIB's because their cards are 2 FPS faster than the competition in reviews...
With NVIDIA cards, there are too many variables anyway, so not all cards boost the same.

Regardless, I'm pretty sure MSI and a number of others AIB partners go through their cards and select some of their best for reviewers to sample.

I've had my hands on two reviews samples - 7970GHz and 7790, and both did over 1300MHz on the core stable. Coincidence?

 

[sm625]

But do they lower the clocks when they do the power consumption tests? That would be really shady.

 

[Stuka87]

I like how his name is repeated constantly, as to like say "I'm so clever everyone look at me". Which in truth none of it really matters

The same with the 970 memory, it was a non issue.

A non issue? Soc heating in reviews, giving a false sense of performance is ok to you?

And nVidia repeatedly lying about the 970 doesn't bother you?

We RELY on reviews to be accurate to the product that we are wanting to buy. What we are seeing here is completely unacceptable.

 

[digitaldurandal]

I like how his name is repeated constantly, as to like say "I'm so clever everyone look at me". Which in truth none of it really matters

The same with the 970 memory, it was a non issue.

Did you forget to put a /s?

You think it is fine to intentionally mislead the consumer? There have been class action lawsuits against larger companies for this type of behavior. It is not a question of whether or not the 1-2% improvement in benchmarks as much as the intent of the action that is the problem.

Also, 970 memory is an issue it just doesn't show up in many titles.

 

[dark zero]

Seems that reviews from the mass media are slowly being untrusting... so the small media is still trusting?

 

[cen1]

Well, I guess I know which brands to avoid on my next purchase then

 

[VirtualLarry]

sooon only press will get a full product and retail will get something different. lol

You mean like with the Kingston V300 SSD?

Anyways, this sort of thing has been going on for YEARS. Remember when mobo reviews, would post performance benchmarks between benchmark brands? Yeah, Asus almost always came out ahead... due to a "minor adjustment" to the clock signal (1-2% OC). Just enough to always come out ahead in the benchmarks.

I remember over 10 years ago it was discovered Asus did the same thing with their motherboards. They were a couple of MHz overclocking the CPU so their mobo was always the best in benchmarks.

Yeah... that.

 

[Bacon1]

press only driver, press only bios. sooon only press will get a full product and retail will get something different. lol

we need better journalists. current once are just pathetic.

Yep :\

 

[Kuiva maa]

Asus is my go to vendor for video cards. Needless to say, I am very ,very disappointed with what I am seeing here. As for MSI...

 

[master_shake_]

this isn't even new.

https://hardforum.com/threads/msi-n...reviewers-retail-versions-are-gimped.1867610/

msi gave reviewers cards with 120 power limit while retail was 109.

horrible practice

 

[MrTeal]

You mean like with the Kingston V300 SSD?

Anyways, this sort of thing has been going on for YEARS. Remember when mobo reviews, would post performance benchmarks between benchmark brands? Yeah, Asus almost always came out ahead... due to a "minor adjustment" to the clock signal (1-2% OC). Just enough to always come out ahead in the benchmarks.



Yeah... that.

A little overclock in the MB space is actually pretty benign compared to some of the stuff they pull with silent revisions in the motherboard space. Reading a review and seeing pictures on the merchant website of one board and then opening one with a couple fewer phases is a little disappointing. Gigabyte (and as you said Kingston) have been as guilty of review unit and unannounced revisions as MSI and Asus.

Caveat Emptor, as always.

 

[Elixer]

The real problem here is that "review" sites don't really care.
They just want the free gear, they could care less if there are any facts in the "reviews". Whatever gets them the click is their motto.

That is why you have to wait for actual consumers to get their hands on the cards (or anything else), and then see how each product does
As we all know, the 970 fiasco got past ALL "reviewers", every single one of them bought it hook, line & sinker, then when the problem came forth by end-users, most of the "review" sites pushed it aside, since they still want their free gear.

 

[tviceman]

I'm going to lose tons of sleep tonight over 2%, even though I manually overclock all my cards well past the maximum specs the card can come with out of the box.