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

[Fastx]

I just read that 290x/780ti H review now per TV and I thought is was a good read imo.

http://www.hardocp.com/article/2014...0x_gaming_4g_video_card_review/3#.UzxeVOmPKUl

 

[VulgarDisplay]

I just read that 290x/780ti H review now per TV and I thought is was a good read imo.

http://www.hardocp.com/article/2014...0x_gaming_4g_video_card_review/3#.UzxeVOmPKUl

That article is pure fail on [H]'s part. They used 14.2 beta drivers which are unfortunately broken. Powertune is broken and the 290x throttles because of this. It's the reason they got such a crappy overclock with their 290x, and the reason why it performed so poorly. Clocks were surely dropping below 1ghz because of the throttling.

Look at the power usage at the end of the article. That should tell you something is wrong right there. The 780ti using more power than an overclocked 290x? Get real haha.

 

[tviceman]

That article is pure fail on [H]'s part. They used 14.2 beta drivers which are unfortunately broken. Powertune is broken and the 290x throttles because of this. It's the reason they got such a crappy overclock with their 290x, and the reason why it performed so poorly. Clocks were surely dropping below 1ghz because of the throttling.

http://www.hardocp.com/article/2014...0x_gaming_4g_video_card_review/9#.UzxIJ_nZS30

Max temps reached were 75 C for the 290x. Did you "pure fail" to read the article? Pretty sure it wasn't throttling.

EDIT: The 780 TI overclock was meh as well. 1120mhz boost is nothing to write home about.

 

[VulgarDisplay]

http://www.hardocp.com/article/2014...0x_gaming_4g_video_card_review/9#.UzxIJ_nZS30

Max temps reached were 75 C for the 290x. Did you pure fail to read the article? Pretty sure it wasn't throttling.

If you can't increase the power limit on the card it throttles. It can throttle form temps, or hitting a hard wattage wall. These cards hit that wall if you don't increase the power limit, and 14.x drivers broke power limit. My r9 290 can't keep 1050mhz at stock settings on 14.x drivers and it's at 65C. If they increased their VDDC offset it forces the card to hit the power wall even faster than if they didn't add any voltage at all.

If I go to 13.12 drivers I can get 1200mhz core easily, and never throttle.

A lot of people at OCN are reporting that if you raise your offset to anything over +40 that the card will throttle even at stock clocks. [H] had theirs at +50. I can't comprehend how they didn't know this about the 14.x drivers.

 

[tviceman]

If you can't increase the power limit on the card it throttles. It can throttle form temps, or hitting a hard wattage wall. These cards hit that wall if you don't increase the power limit, and 14.x drivers broke power limit. My r9 290 can't keep 1050mhz at stock settings on 14.x drivers and it's at 65C. If they increased their VDDC offset it forces the card to hit the power wall even faster than if they didn't add any voltage at all.

If I go to 13.12 drivers I can get 1200mhz core easily, and never throttle.

I think you're just making excuses now for seeing results you don't like. Sounds like they adequately tested their overclock. http://www.hardocp.com/article/2014...0x_gaming_4g_video_card_review/3#.UzxJgPnZS30

"For our overclocking efforts, we set the power limit to +50% (the maximum allowed) and the voltage to +50mV and started to dial up the clock speeds. We began by adjusting the Core Clock clock within Afterburner until we began experiencing instability (note that we had to unlock the voltage modification option within Afterburner and reboot in order to gain access to that functionality). We were able to do some gaming with the base clock set to 1160MHz, however it was not very stable at that speed. We backed down on the GPU speed until it stabilized at 1130MHz. After the GPU overclock was set, we began increasing the memory speed until stability was lost. We were able to achieve an effective 5.4GHz GDDR5 memory rate that was fully stable, even though we were able to run for a short period of time as high as 6GHz GDDR memory speeds."

70mhz less than your overclock isn't going to make a huge difference, anyways. And as I already pointed out, the 780 TI had a lower overclock too.

The point remains - the 780 TI has increased it's lead over the r290x since it was released. BTW, I LOL'd when you said "[AMD DRIVERS]... are broken"

 

[blackened23]

That article is pure fail on [H]'s part. They used 14.2 beta drivers which are unfortunately broken. Powertune is broken and the 290x throttles because of this. It's the reason they got such a crappy overclock with their 290x, and the reason why it performed so poorly. Clocks were surely dropping below 1ghz because of the throttling.

Look at the power usage at the end of the article. That should tell you something is wrong right there. The 780ti using more power than an overclocked 290x? Get real haha.

*rollssssssssssssss eyes*

Sees review that isn't likable. Obviously the reviewer must not know what he's doing, despite same reviewer netting praise just a couple of months back. It's always the fault of the review website. Right? It's a good thing the reviewers at H use the afterburner overlay to see the clockspeed rates in real time and generally aren't stupid about validating overclocks. Of course, feel free to post over at H in their 290 gamer thread. Go ask him. I'm pretty sure he will respond to your claims if you think that he has no clue what he's doing.

 

[VulgarDisplay]

*rollssssssssssssss eyes*

Sees review that isn't likable. Obviously the reviewer must not know what he's doing, despite same reviewer netting praise just a couple of months back. It's always the fault of the review website. Right?

*whatever*

*rollssssssssssssss eyes*

http://www.overclock.net/t/1475034/amd-catalyst-14-3-v1-0-beta-for-windows#post_21965450

Nope. Power Limit is still doing ab-so-lute-ly nothing.

Card throttles to under 900Mhz when putting load on it with +100mV voltage and +50% power limit.
Even at just +31mV and max power limit card throttles to ~1020Mhz under load while it's set to 1100Mhz.

Also, black screening and heavy flickering / artifacting with otherwise stable memory clocks.


Back to 13.12 again...

FIX THE POWER LIMIT FFS. this is a overclocking forum and we can't overclock even 50Mhz like this without throttling!

Care to try again? Feel free to continue trying to troll because you got mad when I called you out for not citing nvidia PR in your posts though.

I'm not saying that 780ti is slower than the 290x. I'm saying they screwed up their review, and probably don't even realize it. The 780ti is the fastest single card you can buy, and the version in this review didn't overclock well. That doesn't change the fact that their 290x had to be throttling because they used a broken driver. Do you all honestly think that an OC'd 290x drew less power than an 780ti? Come on...the card was throttling.

 

[blackened23]

Well then go ask him instead of perpetuating silly conspiracy theories here. I personally think you're just upset at the results. GK110 has already shown a heavy advantage over Hawaii in terms of overclocking. In fact, in the last 290X vs GK110 showdown, the reference max overclocked 780ti was significantly faster than the 290X aftermarket DC II.

Do the math. Two factory overclocked cards. Being that both are overclocked, and with recent NV driver optimizations, the GK110 wins. That isn't to say the 290X doesn't have merit. It's a good card for the price.

But you just need to stop perpetuating this claim of yours. They monitor clockspeeds while testing. If you feel he's testing wrong, GO ASK HIM. If you post in the HardOCP forum thread and ask the editors, they WILL respond to you. So if you think he's an idiot. Stop posting what you're posting. Go to the source. Ask. Then get back to us. I'm confident the results are what they should be. But you're obviously not. So. Do the above.

I'll make it easy:

http://hardforum.com/showthread.php?t=1813287

There's your link to get proof. If you choose to do so. Ask the reviewer if the card throttled. He will tell you if it did. If you don't ask him, then excuse me if I have a hard time taking any of this at face value. If you don't' ask him, then you shouldn't even make another post about it.

 

[VulgarDisplay]

Dear god. Give it up I KNOW THE 780TI is faster.

That does not change the fact that the 290x in that [H] review used a broken driver that caused it to throttle.

I have an r9 290 and it exhibits the same behavior which I noticed when I was overclocking on 14.x drivers. When I went looking for answers to why my 290 was throttling at 65C I found all the posters on OCN that had the exact same problem.

 

[tviceman]

Vulgar LOOK AT the review. LOOK AT IT. Nowhere in the graphs does the 290x show a sudden and permanent drop. [H] tested the overclock, I'm sure they would have observed drops in clocks and throttling (at 75 C ) if it were happening.

You are just clinging onto your own interpretation of why something isn't what you think it should be like.

 

[blackened23]

Dear god. Give it up I KNOW THE 780TI is faster.

That does not change the fact that the 290x in that [H] review used a broken driver that caused it to throttle.

I have an r9 290 and it exhibits the same behavior which I noticed when I was overclocking on 14.x drivers. When I went looking for answers to why my 290 was throttling at 65C I found all the posters on OCN that had the exact same problem.

So you're not going to verify with the reviewer that the card did not in fact throttle. I got it. They monitor clockspeeds while testing. Both Brent and David have stated this in prior reviews many times. If a card throttles THEY KNOW.

If you don't want to get proof for your claims, then nobody can take you seriously. If you ask the editors at HardOCP, THEY WILL RESPOND TO YOU. I promise. They're very open about communication. However, apparently, you're refusing to do so. Why?

 

[VulgarDisplay]

Vulgar LOOK AT the review. LOOK AT IT. Nowhere in the graphs does the 290x show a sudden and permanent drop. [H] tested the overclock, I'm sure they would have observed drops in clocks and throttling (at 75 C ) if it were happening.

You are just clinging onto your own interpretation of why something isn't what you think it should be like.

I'm going to start another thread. I need to reinstall some monitoring software and I will show you all the issue.

 

[blackened23]

I'm going to start another thread. I need to reinstall some monitoring software and I will show you all the issue.

But you're not going to ask the editors at HardOCP. Okay then.

They monitor clockspeeds, as i've said like 4 times now. But will you verify that with them? That's the million dollar question here. It sounds to me like your'e refusing to do so. Are you refusing to do so?

 

[tviceman]

I'm going to start another thread. I need to reinstall some monitoring software and I will show you all the issue.

Do you have the exact same card as (MSI gaming) as [H]? If not, then you are wasting your time. 200 times easier to just ask them to verify clocks while in game. You are just continuing to skirt around the results they got by trying to make up excuses.

 

[tviceman]

Don't worry, I took the 47 seconds out of my day to ask [H] to comment on this thread and it's discussion of their "pure fail" review. Unless it's the answer Vulgar wants to hear, though, I'm sure it'll just upset him more.

 

[Riek]

Vulgar LOOK AT the review. LOOK AT IT. Nowhere in the graphs does the 290x show a sudden and permanent drop. [H] tested the overclock, I'm sure they would have observed drops in clocks and throttling (at 75 C ) if it were happening.

You are just clinging onto your own interpretation of why something isn't what you think it should be like.

Not sure if vulgar statement is correct, but when i look at the review first thing i notice is that an oced 290x is slower than the non oced version in battlefield 4.
All the other result indicate a serious scaling bottleneck (on both cards)
Would be interesting if they would have shown the actual clockspeed in game like anandtech does (sometimes)

 

[blackened23]

Don't worry, I took the 47 seconds out of my day to ask [H] to comment on this thread and it's discussion of their "pure fail" review. Unless it's the answer Vulgar wants to hear, though, I'm sure it'll just upset him more.

Nice. I just emailed him as well. Looking forward to the response.

 

[VulgarDisplay]

I'll ask them as a part of the thread I'm about to post. Interesting findings so far.

Enough derailing the 20nm nvidia thread. Respond to me in my new thread when it gets posted.

 

[96Firebird]

Not sure if vulgar statement is correct, but when i look at the review first thing i notice is that an oced 290x is slower than the non oced version in battlefield 4.

I noticed this, but the text claims otherwise. They mixed the numbers up somewhere...

Overclocking the MSI Radeon 290X GAMING 4G increased performance by 3.7%.

 

[rtsurfer]

Gents while we are at it, how about someone here start a 290/290x vs 780/780ti/Titan benchoff thread..??

Bench a few games at a few pre selected presets. .??

I am pretty sure there are quite a few owners of cards from both sides, if they are willing to take out some time to bench.
We can have our own dataset & draw our own conclusions.

 

[Fastx]

VD I thought it was a good read in showing both oc 1120/1130, and as you see in (example) TR, Crysis 3 graphs no throttling is evident with the 14.2's they used, it looks fine imo. I also thought the BF4 results (FPS) were interesting imo and also those BF4 FPS were with the 290x (Mantle) using 2x MSAA and the 780ti using 4x MSAA. If the 780 ti would have been tested at 2x MSAA also the FPS would even be higher imo. I would also think if H detected throttling they would have mentioned it.

[ Edit] Just for the record I didn't start this thread. I responding to a post by Tviceman who posted about the above H review in another thread and the mods picked my comment and all the above that followed out of that thread (per OT) and moved it here as a new seprate thread which is fine with me.

 

[Attic]

Something was off with the MSI 290x they reviewed in this article. Vulgar's theory appears likely IMO. Throttle makes the most sense for the lower than expected performance on the 290x. We know the 14.2's have this issue, and these are what was used for the review.

The 290x in this review was substantially slower at equal settings than cards they reviewed in the past. A guy in the discussion at Hard noticed this.

Sometimes Brent follows up with this forum, maybe he has an explanation or can confirm if the card is throttling or not.

 

[VulgarDisplay]

VD I thought it was a good read in showing both oc 1120/1130, and as you see in (example) TR, Crysis 3 graphs no throttling is evident with the 14.2's they used, it looks fine imo. I also thought the BF4 results (FPS) were interesting imo and also those BF4 FPS were with the 290x (Mantle) using 2x MSAA and the 780ti using 4x MSAA. If the 780 ti would have been tested at 2x MSAA also the FPS would even be higher imo. I would also think if H detected throttling they would have mentioned it.

[ Edit] Just for the record I didn't start this thread. I responding to a post by Tviceman who posted about the above H review in another thread and the mods picked my comment and all the above that followed out of that thread (per OT) and moved it here as a new seprate thread which is fine with me.

The 1020/1030 benchmarks would not throttle from clock speed, but from the +50 mv they added to the card. In my other thread I showed how I can make my stock r9 290 throttle just by adding voltage. The amount needed of extra voltage to throttle the card due to power tune staying at 0 in all 14.x drivers will vary greatly from chip to chip.

I really wish someone with a 290x would run some tests also because their chip will draw more power than mine causing it to throttle at a lower offset.

 

[tviceman]

The 1020/1030 benchmarks would not throttle from clock speed, but from the +50 mv they added to the card. In my other thread I showed how I can make my stock r9 290 throttle just by adding voltage. The amount needed of extra voltage to throttle the card due to power tune staying at 0 in all 14.x drivers will vary greatly from chip to chip.

I really wish someone with a 290x would run some tests also because their chip will draw more power than mine causing it to throttle at a lower offset.

The problem is that, aside from the 1 BF4 graph (which may very well be a mix up of numbers on the part of the author), the OC still outperforms the stock results on every game they tested. In order for the card to be throttling, it would have to be throttling at both stock and overclocked, yet the throttle when OC would still not be as severe as the throttle when not overclocked? It just doesn't make sense to blame the review on throttling. Reverse the numbers on the single apples-to-apples BF4 graph and everything suddenly checks out. The OC'd results beating stock across the board, the performance scaling relative to the overclock, and the power consumption of stock vs. overclocked....

Yeah, no throttle going on here, other than the 780 TI throttling the 290x.

 

[VulgarDisplay]

The problem is that, aside from the 1 BF4 graph (which may very well be a mix up of numbers on the part of the author), the OC still outperforms the stock results on every game they tested. In order for the card to be throttling, it would have to be throttling at both stock and overclocked, yet the throttle when OC would still not be as severe as the throttle when not overclocked?

It just doesn't make sense. Reverse the numbers on the apples-to-apples BF4 graph and everything checks out. The OC'd results beating stock across the board, the scaling, the power consumption of stock vs. overclocked....

Yeah, no throttle going on here, other than the 780 TI throttling the 290x.

You are assuming that when the the over clocked card is throttling that its clocks are doing two things: dropping lower than the stock card speed, and when it throttles consistently staying lower than the stock card speed. The core clock will still stay higher than the stock core clock, and depending on how demanding the scene will often return to the advertised oc core clock as demand drops making the card draw less wattage.

A throttling over clocked card will still be faster than the stock card, but will not put in the best benchmark run possible if it wasnt throttling.

Let me make it clear that I in no way expect the 290x to be faster than the 780ti. Just that [h] may have unwittingly lowered the 290x scores somewhat due to their settings.