Home / News / Desktop PC News / No, Asus and MSI haven't been tricking gamers with tweaked graphics cards

No, Asus and MSI haven't been tricking gamers with tweaked graphics cards

by

GTX 1080 hero

When it comes to graphics cards, benchmarking stats are the single most important consideration.

Which is why Asus and MSI find themselves at the centre of some controversy after both companies were accused of sending out overpowered GeForce GTX 1080 samples to reviewers.

But both companies have responded to the claims by explaining that review models are sent out in a higher performance setting that retail customers are also able to, and probably will, access.

The controversy began when TechPowerUp noticed a review model of the MSI GeForce GTX 1080 Gaming X running at faster GPU and memory clock speeds than the retail version.

As it turns out, the card was set to use OC, or overclocking mode, out of the box, as opposed to the standard Gaming mode that retail units are set to.

Related: Best Graphics Cards

To be clear, regular consumers can also set their cards to the slightly more powerful OC mode simply by changing a setting in the special software that comes with the cards themselves.

But the report from TechPowerUp claimed that using OC mode as the default setting in review samples is "not 100% representative of retail cards, and is questionable tactics".

The concern is that benchmark numbers from OC mode could be represented as standard performance figures, which when compared with the other cards on the market, would give the company a slight edge over competitors.

But as standard practice should be to test third party cards on both standard and overclocked settings, it seems surprising that this is only now being highlighted as a concern.

Still, Asus was also accused of adopting the practice in the report.

Related: Best Gaming PC – The ultimate PC building guide

MSI has previously responded to the issue, providing a comment to HardOCP Editor-in-Chief Kyle Bennet in 2014:

"We can assure you that this is not some sort of shenanigans or a marketing ploy. We did send the media review samples with the OC Mode speed selected which is included in the program called the “Gaming App.” This setting is different than what the retail video cards are being shipped with.

"All consumers who purchase our “GAMING” VGA cards should have the default Gaming Mode clock speed selected, but by installing the “Gaming App” you can increase the clock speed of what was represented in most reviews by pressing the OC Mode button. This does NOT void the warranty.

"We simply wanted to promote our software called the “Gaming App” which is only included in our GAMING branded video card products. We sincerely apologize for the inconvenience for not explaining it clearly to some of the media that published reviews and to the gamers and enthusiasts that purchased one of our GAMING products."

And Asus has provided an official statement in response to the accusations, claiming that setting review samples to OC mode by default "saves time and effort" for reviewers

Here's the full statement:

"ASUS ROG Strix GeForce GTX 1080 and GTX 1070 graphics cards come with exclusive GPU Tweak II software, that provides silent, gaming, and OC modes allowing users to select a performance profile that suits their requirements. Users can apply these modes easily from within GPU Tweak II."

"The press samples for the ASUS ROG Strix GeForce GTX 1080 OC and ASUS ROG Strix GeForce GTX 1070 OC cards are set to “OC Mode” by default. To save media time and effort, OC mode is enabled by default as we are well aware our graphics cards will be reviewed primarily on maximum performance. And when in OC mode, we can showcase both the maximum performance and the effectiveness of our cooling solution.

"Retail products are in “Gaming Mode” by default, which allows gamers to experience the optimal balance between performance and silent operation. We encourage end-users to try GPU Tweak II and adjust between the available modes, to find the best mode according to personal needs or preferences.

"For both the press samples and retail cards, all these modes can be selected through the GPU Tweak II software. There are no differences between the samples we sent out to media and the retail channels in terms of hardware and performance."

VIDEO: Catch up on all the latest tech news with The Refresh

Have Asus and MSI failed to communicate the difference between retail and review models of their graphics cards? Or is this a legitimate practice? Let us know what you think in the comments.

Phil

June 21, 2016, 5:40 pm

I'm sorry but this is bullshit. You're defending the indefensible. I can overclock my CPU as well using "special software" - doesn't mean that a reviewer should test 3GHz as "standard" when I get a 2.5GHz CPU out of the box. What comes to the reviewer should be the same product in every way as I get off the shelf. "Saves time for reviewers"??? Really? I'd have thought it would mean any honest reviewer would have to WASTE time putting the thing back to stock settings. It just means the less thorough reviewers who don't notice and just benchmark what comes in the packet make their products look better. This is disgusting tactic - software is integral to performance just as much as hardware. If they sent you different hardware under teh same name with a user replaceable chip and said "well, you can alter the hardware back to stock without voiding the warranty" would you be happy? Honestly can't believe you're defending this and I can only assume it's so they will continue to send review samples and invite to events.

MattMe

June 22, 2016, 11:48 am

The state of it all is a little worrying, isn't it?

Michael Passingham

June 22, 2016, 12:29 pm

Hi Phil, I didn't write this piece but as the guy who does Trusted's GPU reviews, I thought I'd share my thoughts.

To an extent, I get what you're saying. It's at most slightly cheeky to ship a card pre-clocked for top performance, but at the same time it should be the media's job to test a GPU in all circumstances, and that includes downloading all relevant software to test performance under different conditions.

If we're not doing that, we're not doing anything beyond what a consumer with a couple of games and Fraps can't.

What this whole thing highlights is that these sites, whoever it may be, need to pay closer attention. I'm amazed not everybody isn't using GPU-Z or an equivalent to double check specs before barreling into a review.

Reviews are results tested against claims. If you don't have the full set of results, you can't say you've tested their claims fully.

I've spotted so many mistakes in many spec sheets for products from huge manufacturers (of all sorts of hardware, not just GPUs) that I can't take a manufacturer's word as truth until everything has been double checked. I feel like I spend half life my correcting info that's come from specification listings that are outright wrong.

Mistakes are made, and the media should be there to correct them and tell the whole story.

If there's overclocking software available for a graphics card, a reviewer should be installing it, otherwise they're not telling the whole story in their benchmarks and not serving readers a complete review.

While I don't think the parties involved went in trying to mislead, I would expect any card manufacturer worth its salt to provide a reviewers' guide, and in my experience with new products, this is almost always the case. If they're not making this absolutely clear upfront, then they need to start to avoid this in future.

On your final point about receiving review products and event invites, that's not in any way true. With components in particular, we have many ways in which to source them beyond direct contact with a manufacturer (such as through a retailer), so we're happy to be highly critical where it's due.

Phil

June 22, 2016, 1:27 pm

Thanks for the reply. It's nice to know you are indeed looking at things in depth and not taking their word for it. What I want to know though is when you're posting benchmarks for a GPU (the standard ones, not OC ones) are you resetting the clocks to the standard speed as we the consumer receive them or leaving them in their higher clocked states as the manufacturer supplies them to you?

It's also reassuring to know that you don't just rely on units from manufacturers to allow you to be more honest. I'm sure you can see why people will think that a negative review might result in you being disadvantaged by a manufacturer - just look at what happened to Anandtech with OCZ's SSDs. He pointed out poor random performance and that this had been sacrificed to gain peak sequential (for marketing) at a massive cost in usability, refused to recommend the drive and they stopped sending him review units for a time.

Michael Passingham

June 22, 2016, 1:47 pm

So far at Trusted I've not reviewed any third-party cards, only Nvidia's ones. Nvidia ships their cards as they would to a consumer.

However, it's always been my policy (in previous jobs as well as this one) to test any given component in whatever modes the manufacturer make possible, and installing all manufacturer software before embarking on a review. I'd urge all other reviewers to be as transparent as possible (and most of them are) about this so readers/consumers continue to trust reviewers as a whole.

Evan

June 22, 2016, 2:04 pm

Hi Phil, I'll let Michael answer your question but I'd like to cover your second point.

It's true that we sometimes don't get as open access to products as others simply because of the fact that we review products thoroughly and are honest.

Thankfully our sheer scale means we can't easily be ignored. On occasion I have also simply purchased products that haven't been sent to us, or, as Michael mentioned, sourced from retailers.

TrustedReviews has been around since 2003 and the ethos that it started with hasn't changed when it comes to reviews – they must be clear, honest, comprehensive and impartial. Pandering does not build a sustainable business, readers are too savvy and we value them, and our credibility, too much.

Manufacturers often don't like comparisons with other products but we do that in every review and regularly tell readers that another product might be a better fit for their needs. They certainly don't like that.

What annoyed us with this particular story was that it was blown way out of proportion. This is not like when Samsung and HTC boosted their phone chips when benchmarks were being run, or the Volkswagen emissions fiasco – both stories we covered.

Phil

June 22, 2016, 2:20 pm

Might I make a proposal? In order to reassure people that what you're testing and quoting is what they'll get off the shelf, when you're producing the benchmark graph state what clock speeds were used or simply put "stock speeds" at the top when you're testing at the settings we get off the shelf. I make a decision on upgrading by looking at what I'm getting from my GPU in a game and then comparing that to the numbers produced in a review - there are lots of variables but this allows me to see if it's worth upgrading or not and so if they're being tested at different speeds it will seriously impact my buying decision.

It's true that I've been reading TR for quite some time, mostly on the recommendation of integrity and lack of pandering to manufacturers. Whilst you might well test cards thoroughly it does worry me that in a lot of these kinds of jobs, especially physical magazines who have been hit hard by the rise of internet review sites, time pressures mean they might well simply test what's in front of them as quickly as possible without the same level of scrutiny being possible. I thoroughly believe that whatever you get should be set up identically to how the consumer receives it.

Michael Passingham

June 22, 2016, 2:29 pm

I think based on this story and the amount of traction it has had, we'll have no choice but to be ultra clear on exactly how things have been tested. It won't actually change how we do things (we've been doing it properly) but since the reader might now be a little more suspicious of GPU tests, it's probably a good idea.

Phil

June 22, 2016, 3:29 pm

Well I'm glad you guys have been testing things properly and whilst I accept you're in the industry and know how things work and how products are tested far better than I do, I'm sure you can see how this might make the average reader think that companies are trying to trick unscrupulous / time pressured reviewers. Thanks for taking the time to respond to my (quite angry) rant. It's good to see a website engaging with its readership.

comments powered by Disqus