MSI and ASUS Send VGA Review Samples with Higher Clocks than Retail Cards.

http://www.techpowerup.com/223440/m...-samples-with-higher-clocks-than-retail-cards

>_>

 

I'm glad they are getting called out for this. It's bull**** and it shouldn't happen.

Edit: The frequency differences are pretty low, it probably wouldn't even be visible in game. Also apparently MSI advertises that speed on their site. That being said it still shouldn't happen.

 

1fps is life or death, true story bro.

 

It's not so much that, it just sets a slippery slope of them modifying other things, like fan profiles, binned chips specifically for reviewers, etc. It just shouldn't happen. I don't really care what their excuse is, the review should be what the card is when I pull it out of the box.

 

This s*** has been going on like forever. The only true reviews are ones where the reviewer is not affiliated with anyone related to the card.

Seriously though, if a product is the same price, everything else is the same, but, is 1fps faster, you telling me you're not going to buy that 1fps faster product?

I bet you 99% would for sure. Which is why these cheating tactics happen. We, the consumers are TOO savvy.

 

Thanks for clarifying the situation HH and your honest reviews.

But, I beg to differ that it doesn't do much especially as I've noticed with each new gen; small overclocks are more noticeable. Just use Furmark, I guarantee you will see fps difference with just 25mhz oc.

3dmark score? extra 1fps here and there in games? Certainly, enough to put their product at the top of your charts to make us think they're the brand to go for. It plays with the mind of us consumers and if it's helped them get more sales, then, that 25mhz has done it's job. Costs nothing and is free advertising when given to reviewers like yourself.

The fact there was another "final" bios available as soon as you called them out on it should tell you something. Both Asus and MSI are excellent brands, but, these dirty tactics need to be called out.

 

Thanks for explaining and reviewing like a boss!

This 25MHz correction may not be much, but in consumer minds, it's a lot when you have 2 brands at the same price and one is marginally faster. It would make no sense to go for the slower card in this scenario, even if it was 100 vs 101 fps.

 

eh....

I know it's not 100% across the board like this, but 25mhz would mean nothing. 1683 to 1709 is an increase of 1.533%...which means you shouldn't get an increase of 1fps until close to 100fps.

For benchmarking, this would be seen as a variable that is known to happen between each test and would mean nothing. Not only that, but with benchmarking most games are looked at for their highest graphics at high resolutions and likely never get to 100fps anyways, at least for what people are looking at mostly.

Again, i know that a 1.5333% increase in frequency does not necessarily mean a 1.5333% in FPS, in fact it's likely lower, so really the "best case scenario" doesn't even mean anything

 

Mine has the 1683MHz clock.

Well, like you say, the difference is negligibel and the issue is non-existent since cards clock very differently from what I gather.

Still it's bad PR to "cheat" like that and this behavior should not be endorsed.

 

Can't say I'm surprised. Review cards themselves are guaranteed to be cherrypicked.

 

It is a minimal difference, but, if you say "ok to 25mhz", then, at what point do you draw the line?

Are you going to say 25mhz is nothing, but, 50mhz is a "no no"?

These differences are certainly easier to see at lower resolutions like 1080p, which consumers will look at.

Just try MSI's Furmark/Torus and overclock your GTX970 by 25mhz and see if it changes the fps. I guarantee you will see higher fps.

 

What I'm more worried about, is the other possible changes.

Its not just the boost clocks and default clocks that can be changed, a slight voltage increase is possible, giving it slightly more headroom in max clocks.
It could even bring the maximum power consumption/maximum allowed voltage higher than the retail cards.
We just don't know.

That why we have to rely on the cards being reviewed as being representative of what we can purchase.

Hilbert shouldn't have to email MSI and Asus to tell them they screwed up.
According to TPU, its not like this has just begun. Its been going on for a while.

Hardware.fr have been asking for a commercial_BIOS, and were told multiple replies, ranging from "You would have tested it in default mode only otherwise" and "Its only a few percent, don't pay attention to it...".
These are replies from large companies, who should know better by now.
They also state that some of the BIOS's, when de-clocked to default, aren't actually back at default.
They had to force the card into thermal overload to read the correct base clock value, which was still overclocked.

As they say in closing up, now its 25MHz, or 1 or 2%.
But when its 3 or 5%, do we just give up ?

 

This^^ When the cards being sent to hardware sites are cherrypicked golden samples, we already know that the results are going to be different from the results we are going to get with cards we buy in the stores...

It´s lame and shouldn´t be done, but every company is gonna use small tricks to make their products look better so we buy them...

 

Just been reading the comments on Hardware.fr, and one of the site reviewers came up with this information about some of the differences between the two MSI Bios's on the GTX 1080 Gaming X (Italics is what I have translated) :

"bios reviewer : 1708/1847/1264 MHz with power limit of +/- 250W
bios that solves the "problème" : 1683/1822/1251 MHz with power limit of +/- 220W"

So the power limit was raised also, which will make a large difference in the actual Boost clocks, on demanding programs/games.
He does state that on the Witcher 3, the drop going to the retail BIOS only dropped 26MHz, but on FurMark, it lost over 100MHz.

edit2 : Asus only changed the 26MHz, they didn't modify the Power limit, so the only difference was the 26MHz, nothing else.

 

I don't think cherrypicking is working any more.

As the fab process goes down in size, we seem to have less and less headroom.
Maybe its due to the architecture, i'm not entirely sure.

All of the 10x0 seem to plateau at 2.0-2.1GHz MAX.
Even on water you're not going to see much, if any, more.

i have a 200W peltier that I never got round to using, and don't think I will ever be able to use it in the future...apart for the car (Home made A/C)

 

Well, from my point of view, this is far from a screw-up or mistake.

This was deliberate. This targeted reviewers which in turn target us, the consumers.

 

Agreed. A slightly different clock I can live with, different max volts or other things not so much.

 

If consumers are so easily lead by their nose; letting such results as this 1fps difference determine which card to purchase, is so sad. Why on earth would someone go with a brand they have no experience with because it is 1fps faster? It is easier to go with a brand that one has had good experiences with and just oc it to match the said superior performance of the other card.

The whole silicon lottery should render such close results being a determining factor null and void.

 

The dirty thing is that i might buy a 1080 just because it's the fastest even if it's only a tiny fraction.

These things are what help people to make a decision on which product they are going to pay a lot of money for.

Even though it is a tiny amount people will be put off by this deception whether it was intentional or not.

Luckily for us here HH isn't fooled by these dirty tricks.