[H] MSI Radeon R9 290X Gaming 4G Video Card Review

[Schro]

@Schro
If you guys bench things differently between runs and person doing the testing, it would lead to huge variations which we've seen of late in your reviews.

For example, person A will do a run through at a particle area of a level, that is either more or less punishing on GPUs compared to person B.

Why do you guys not standardize the run and play through? Picking a random section is going to skew the result and not representative.

The only conclusion from comparing your own reviews:
XFX R290X is magical.
MSI R290X is crap.
More recent AMD drivers lose massive performance.

Which is it?

We do have a standard runthrough that each of us uses for each of the games. Each runthrough will start in the same place every time and work through a level until a certain point. The end of the runthroughs can change over time depending on the relevance to the overall gameplay experience - Crysis being an example - we can determine rather easily in 5 minutes of game play as to the appropriateness of particular best playable settings. While collecting 10 minutes of data can be better, that will add nearly an hour of work to complete the review where no value is added for the reader (as the end result will be same resolution and same system spec). BF4 is very unstandard due to the multiplayer nature of the beast.


Now, for the conclusions, there's a few factors in play. I do not think that the MSI card is crap. I actually think it is a better built card than the ASUS DirectCU II model and it is also quieter (I have not spent time with the XFX one yet, so I can't give you a comparison on that). At the time of publication, the MSI card was priced at $650 which was a $60 premium over the XFX and $30ish over the ASUS. The ASUS has a higher stock clock and memory clock for less money - therefore the MSI one is much less appealing. Of course, since then, prices have fallen, to where after MIR, the MSI card will be on par with the XFX - in that case, I'd probably be more kind to MSI in the conclusion. MSI also lost points with me due to the execution of their GAMING app (and related vision).

 

[Schro]

Obviously you forgot that the [H] ASUS DCII required very high fan speed to prevent it from hitting 94C.. something other sites do not experience, far from it. They had terrible scaling in performance for OC DCII when other sites disagree.

http://www.guru3d.com/articles_pages/asus_radeon_r9_290x_directcuii_oc_review,28.html

So either [H]'s DCII sample was bad (and ASUS/Gigabyte sent out bad samples per their own statements), or Guru and other sites are somehow magical samples.

We have two different DCII samples in house and I've spent a good bit of time with each of them (individually as well as in crossfire). At stock clocks, they perform just fine and exactly as I would expect from that level card. For both of them, when overclocked, power consumption goes through the roof which in turn generates a lot of heat. I noticed that the MSI card did NOT do that.

.....

Sorry - quote system snipped you. I'll do my best to answer here...

Our run-throughs are for the same level. We start with a run-through that Brent assembles and communicates to us. My gameplay style is going to be different than his and anyone else's playstyle, even if there are no other variables. Keep in mind that our goal is NOT to communicate the FPS of something, but to communicate the in-game experience.

As for 290X vs 780 Ti, at reference clocks, they are both fairly equal to each other. When OC'ed to the max, the 780Ti has far more headroom in the silicon and tends to slap the 290X around with a trout. The MSI 780 Ti GAMING card has a 20% factory overclock built into it, so it will win just about any competition with ANY 290X (except for maybe your card).

The review published on Monday. I'm not sure how the numbers could have been wrong since January?

I'm not saying your cards are cherry picked nor am I saying that I have tried to overclock a statistically valid sample of cards. I'm impressed you're getting those numbers out of your card as to date, we have not seen that at [H].

 

[Silverforce11]

We have two different DCII samples in house and I've spent a good bit of time with each of them (individually as well as in crossfire). At stock clocks, they perform just fine and exactly as I would expect from that level card. For both of them, when overclocked, power consumption goes through the roof which in turn generates a lot of heat. I noticed that the MSI card did NOT do that.

The reason I brought that up was that other sites, Guru, TPU, Legit, HC, Kitguru and even AT here have found good scaling when they OC R290/X... except for your ASUS DC2 sample in the prior published article and now this MSI. Your DC2 card also had a weird result in requiring max fanspeed to prevent it from running so hot when other sites did not find that. Specifically TPU and Guru did not need to run high fan speed for a bigger OC and they noted how good the cooling solution was operating.

It's reminiscent of the Computerbase.de result when they had a bad Gigabyte WF3 R290 and ASUS DC2 that actually throttled unless fan speeds were cranked up. Gigabyte also confirmed they sent out a bunch of bad cards.

What we get from reading your review is that the DC2 and now MSI R290X doesn't scale well with OC, but thats contrary to user results as well as other review sites. So what is it? Bad sample? Testing methodology? Just random luck? Because if I took it at face value, I would avoid the DC2 and MSI R290X and buy XFX, PCS or Sapphire brands only.

 

[VulgarDisplay]

Schro, how do you confirm that your cards are not throttling? Are you actively logging for review after your benchmark runs?

 

[Schro]

What we get from reading your review is that the DC2 and now MSI R290X doesn't scale well with OC, but thats contrary to user results as well as other review sites. So what is it? Bad sample? Testing methodology? Just random luck? Because if I took it at face value, I would avoid the DC2 and MSI R290X and buy XFX, PCS or Sapphire brands only.

Can you link some of the other sites reviews specifically so I can look at methodology to see how they compare?

Personally, I would expect to see similar scaling across all brands assuming the test methodology is held consistent across tests. The DC2 review was done by Brent while I did the MSI review, so that does mix things up a little bit.

Schro, how do you confirm that your cards are not throttling? Are you actively logging for review after your benchmark runs?

For me, I tend to have Afterburner running in both the OSD as well as on the LCD of the G15 keyboard I have on the review rig, so I can keep an eye on it that way. I will also go back after each test run and review the Afterburner stats (or GPUz stats, depending) to see if the clock throttled during the run.

 

[Silverforce11]

Fan 50% speed, bigger OC, lower temps.
http://www.guru3d.com/articles_pages/asus_radeon_r9_290x_directcuii_oc_review,28.html

Note the temps and fanspeed, better temps, lower fanspeed. Higher OC.
http://www.hardwarecanucks.com/foru...94-asus-r9-290x-directcu-ii-oc-review-11.html

Default fan profile, scaling just fine with OC.
http://www.techpowerup.com/reviews/MSI/R9_290X_Gaming/27.html

Other sites find even reference cards scale fine with OC when set to uber or higher fan speed. So when you guys test your DC2 and MSI and have terrible scaling, and it runs so hot requiring very high fan speeds, it rings a bell.

Either way, I've been recommending users here and IRL to avoid DCII and WF3 R290s and go for other brands such as Sapphire or Powercolor instead.

 

[Schro]

Fan 50% speed, bigger OC, lower temps.
http://www.guru3d.com/articles_pages/asus_radeon_r9_290x_directcuii_oc_review,28.html

Note the temps and fanspeed, better temps, lower fanspeed. Higher OC.
http://www.hardwarecanucks.com/foru...94-asus-r9-290x-directcu-ii-oc-review-11.html

Default fan profile, scaling just fine with OC.
http://www.techpowerup.com/reviews/MSI/R9_290X_Gaming/27.html

Other sites find even reference cards scale fine with OC when set to uber or higher fan speed. So when you guys test your DC2 and MSI and have terrible scaling, and it runs so hot requiring very high fan speeds, it rings a bell.

Either way, I've been recommending users here and IRL to avoid DCII and WF3 R290s and go for other brands such as Sapphire or Powercolor instead.

It is tough to tell exactly how they test the OC of the card. Finding the maximum stable overclock across a large suite of games is one of the most frustrating parts about doing these evaluations. For example, I had the MSI card running fine at 1080/6000 across a couple of games in our suite, but it then wobbled and fell over. The reductions that you see are the result of that. Before doing data collection at a particular speed, we will spend several hours playing a few different games to make sure the maximum overclock is stable no matter what we throw at the card.

Running a game's canned benchmark or time demo can also skew things. For example, Hitman has a built in benchmark utility, however I could run it over and over and be stable at that clock speed, but then if I loaded a level I'd crash out within a minute. The canned benchmarks can also impact the power usage and heat creation that occurs when overclocking, leading to different observations.

(hopefully that makes some sense or answers your question. I'm falling asleep while typing right now... )

 

[blastingcap]

I have no complaints about my ASUS 290s.

If you have a decent cooling case, not the hotbox Tom's used, a non-Crossfired 290 of any brand, and even the reference model, will do fine. I have a pair of ASUS DCuII's and am fine with them. http://www.bit-tech.net/hardware/graphics/2014/03/03/asus-radeon-r9-290-directcu-ii-oc-review/8 (temps are deltas not absolute, but you see how well the temps are compared to other video cards)

Temps will go up if you try to run two cards in Crossfire next to each other. That requires even better airflow. But all the 290/290X cards, the MSI Gaming and Sapphire Tri-X included, will struggle somewhat with temps in Crossfire due to the nature of fans given so little room to work with and the limited airflow. The top card will be especially toasty. It may not rise to the level of actual throttling, just that the temps will increase more than if you had one card alone.

 

[VulgarDisplay]

I've been watching HWinfo on my r9 290 and it throttles by as much as 70mhz in all games on any 14.x driver. I do not understand how [H]'s 290x did not throttle in their benchmark using these drivers.

The only way to use these drivers and not throttle when overclocked is to flash the asus PT1T bios on your card. As soon as I overclock my ram I can get my card to throttle at stock clocks. The higher your overclock and higher your VDDC offset the more your card throttles. I've also found the 14.x drivers to just be terrible all around overclocking drivers because they aren't completely stable even at stock settings.

 

[Fastx]

I've been watching HWinfo on my r9 290 and it throttles by as much as 70mhz in all games on any 14.x driver. I do not understand how [H]'s 290x did not throttle in their benchmark using these drivers.

The only way to use these drivers and not throttle when overclocked is to flash the asus PT1T bios on your card. As soon as I overclock my ram I can get my card to throttle at stock clocks. The higher your overclock and higher your VDDC offset the more your card throttles. I've also found the 14.x drivers to just be terrible all around overclocking drivers because they aren't completely stable even at stock settings.

VD just thinking here, you bought a Gigabyte GTX 780 around a month ago and posted you were getting artifacts and throttling, now your 290 is also throttling so any chance you have an under powered PS or have a PS going bad?

Also the one guy I linked you to in your other thread says he has no throttling on his MSI 290 Gaming card with 14.2's but did with 14.1's and remember H review was with 14.2's with a MSI 290x Gaming card. Just curious if you happen to have any forum/post links from your research on this problem with a MSI 290/x Gaming card throttling oc at 1120 using 14.2's?

BTW you also posted in your other thread you may or may not see this problem depending on your cards BIOS. So I am taking this as depending on the card brand/BIOS.

 

[VulgarDisplay]

My card only exhibits this behavior on 14.x drivers. No throttling on 13.12 at all so my PSU (750w corsair) is fine.

If I raise my voltage to +37 offset and up my clocks to 1100mhz the card throttles to 1023 on the core and stays there.