Best Graphics Card for Phenom II X4 980BE

[Clutchc]

you were asked if you were running physx in Metro2033 but I never saw you answer. and the bench for both Metro games will hitch a bit unless you have more than 4 threads.

Oh, sorry. I got sidetracked. I turned PhysX off to see if it would be any better. If it was, it was un-noticeable. And I forgot to save the results.

Does anyone have 3DMarkVantage for comparison? I can run that. It has both GPU and CPU scores. I have the reg. ver. so I can run the default settings or custom.

 

[Silverforce11]

What games do you play? Go play it instead of worrying about Metro benches which aren't representative of gameplay.

3dmark has your rig where it should be, the GPU is definitely improving your performance in graphics intensive situations. But if you play some games that are 1 or 2 threads, you may not get much of an improvement, but here's where you can crank up the settings and AA, whereas the 750ti would be crippled under max settings.

 

[Clutchc]

What games do you play? Go play it instead of worrying about Metro benches which aren't representative of gameplay.

3dmark has your rig where it should be, the GPU is definitely improving your performance in graphics intensive situations. But if you play some games that are 1 or 2 threads, you may not get much of an improvement, but here's where you can crank up the settings and AA, whereas the 750ti would be crippled under max settings.

OK. Fair enough.

Mostly FPS. Some TPS. My favorite type are games like the Crysis series. I like games that are rich in detail and rich in story. And allow stealth and tactics instead of run-and-gun.

 

[Termie]

One of the best ways to monitor CPU and CPU usage is the new MSI Afterburner Beta, which includes real-time charts for both. You run that in the background while gaming, and can then see the usage over time. It's extremely handy. I'm not sure if the standard version of Afterburner does that - I've been using the betas for a while and I think they added the CPU charting that the standard version didn't have...not sure.

You can see the link to the beta on this page, after the link to the standard version: http://event.msi.com/vga/afterburner/download.htm

I wouldn't worry about 3dMark Vantage - it's so out-of-date that it won't show any benefit to your video card. And like others have suggested, the Metro2033 benchmark is wonky. Maximum settings on it will bring just about any video card to its knees. I believe I last scored about 27fps in that benchmark on a GTX 670, which is a bit faster than your card. That was "unplayable," but of course it's just a benchmark, you don't play the game at those settings.

 

[lopri]

Something like GTX 670 or HD 7870 would be upper bound for that CPU. I am getting a 290 shortly and can test it on a 1045T. My 1045T is OC'ed quite a bit but it's not difficult to test in stock configuration.

 

[Clutchc]

@ Termie
Thanks for the heads up on the MSI Afterburner. I have the reg ver. I'll have to check if it has that. If not, I'll download the beta.

Now... I wonder if my HDMI issue is worth having Newegg exchange the card for a new one?

 

[Termie]

The HDMI issue is troubling, but it could be a lot of issues unrelated to the video card. You say it goes out when starting up a game, right? Do you have to restart to get the DVI to work, or do you have that plugged in as well and can you switch immediately to it?

Does it ever go out when not gaming? And when it does go out, what happens next? Does the signal come back when you quit the game (perhaps by using control-alt-delete, or do you do a hard reset to get it to come back on.

I'd really like to know if you can switch to DVI while experiencing a blank screen and actually see the game running.

 

[escrow4]

I'd bet $100 that the old Phenom II is holding that card back, especially the minimum's.

 

[Makaveli]

I'd bet $100 that the old Phenom II is holding that card back, especially the minimum's.

Yes I do believe he is cpu bottlenecked also.

For most of the builds i've seen with those cpu's and radeons anything over the 6xxx is usually held back and even with 6xxx radeons you will need a fairly decent overclock to keep them feed with a Phenom II.

 

[Clutchc]

The HDMI issue is troubling, but it could be a lot of issues unrelated to the video card. You say it goes out when starting up a game, right? Do you have to restart to get the DVI to work, or do you have that plugged in as well and can you switch immediately to it?

Does it ever go out when not gaming? And when it does go out, what happens next? Does the signal come back when you quit the game (perhaps by using control-alt-delete, or do you do a hard reset to get it to come back on.

I'd really like to know if you can switch to DVI while experiencing a blank screen and actually see the game running.

The display goes out when I ran Hitman Absolution on HDMI. The other monitor on DVI remained on. Right now I am on DVI for both monitors and haven't had any issues.

No, it never went out at all except for when Hitman started. I can't remember if I got back the display after doing the control-alt-delete thing or not. After the issue, I went to both monitors on DVI and had no more trouble.

Next time I try the HDMI output, I'll try your test. But when the HDMI display went black, the monitor displayed the floating "input not supported" msg. It wasn't just a black screen. That made me wonder if the HDMI signal was out of range. Could be that one game, I guess.

 

[Clutchc]

I DO feel better about the card/CPU combo now. I kept MSI Afterburner on the other screen and monitored the GPU and CPU usage in percent, as well as framerate and frametime.

Crysis 3 @ 1080p, all settings at HIGH.
GPU Usage (%) = 60's to high 90's
CPU Usage (%) = 60's to 70's all 4 cores looked like they were independent and reading relatively similar with some obvious variations here and there.
CPU Core Max Temp = 51C
GPU Max Temp = (I forgot to note the number, but it was safe)
Frametime (ms) = 20's to mid 30's with occasional spikes
Framerate (FPS) = low 40's to mid 50's
Overall game play feel = smooth

Hitman Absolution Benchmark @ Max Settings
Min/Max/Avg (fps) = 35.8/56.3/44.9
CPU Usage = ~80% to ~90% all cores showing independent fluctuations
GPU Usage = ~60% to ~70%

Heaven Benchmark @ max settings: http://postimg.org/image/85x7vv457/
Valley Benchmark @ max settings: http://postimg.org/image/9m8qe0723/
Metro 2033 Benchmark (see settings listed): http://postimg.org/image/ukewc36wr/

So... seem about right for a 3.74ghz Phenom II X4 and R9-280?

 

[toyota]

I dont know where you are testing in Crysis 3 but no way you would stay above 40 fps in many areas. even my stock 2500k would drop into upper 30s. oced to 4.4 I dropped into mid 40s and with my 4770 oced I get right at 60 fps in same spot. this was tested on the first level.

 

[Clutchc]

I dont know where you are testing in Crysis 3 but no way you would stay above 40 fps in many areas. even my stock 2500k would drop into upper 30s. oced to 4.4 I dropped into mid 40s and with my 4770 oced I get right at 60 fps in same spot. this was tested on the first level.

It was in Only Human, I think. I quit when I got to the checkpoint where you free the 2 guys from the crashed VTOL and get the JAW missile launcher and restock ammo.

I just report what MSI Afterburner shows on the graphs. I have no reason to be making this up. I'm here for help, not to brag.

 

[Termie]

You're good. Those measurements seem exactly right, and keep in mind that those are two of the most CPU bottlenecked games.

 

[toyota]

well I am looking at Crysis 3 now using up to 90% of my oced 4770k probably averaging 65%. it makes no sense that a Phenom 2 was not even getting maxed out and he was getting in low 40s for framerate. he must be testing at the least cpu bound part of the game. that game ate my oced 2500k for lunch at times and its was obvious.

EDIT: screen I took from last night


how to print screen on pc

 

[Clutchc]

@ Toyota
I may have missed a few spikes higher than 70%, but whenever I moused over to the other monitor, that's what the numbers were at the high points in the graph. Do you think it's reporting incorrectly? I don't see how I could have read it wrong. I made sure I was in the correct graph. It's the graph below the title heading, right?


@ Termie
Hey, great! I think I'll accept the results then, and either put the B59 back in for further testing... or continue to OC this 965BE. Thanks.

Btw, my other machine is a 2500K @ 4.2GHZ/R9-290X. I attempted to check it against the results as per toyota, but the 2nd monitor is connected HDMI, and the same thing happened that happened with my 280 & Hitman. As soon as the game started on the DVI monitor, the HDMI went black until I quit the game. So, no real-time graph to compare.

 

[Majcric]

I dont know where you are testing in Crysis 3 but no way you would stay above 40 fps in many areas. even my stock 2500k would drop into upper 30s. oced to 4.4 I dropped into mid 40s and with my 4770 oced I get right at 60 fps in same spot. this was tested on the first level.

You're not using Sli obviously from specs. An i5 to an i7 scales that well. The web doesn't show it, but I trust you Toyota.

Sli has really been the only major gains I've seen from an i7 so far....

 

[Clutchc]

My other machine (i5-2500K/R9-290X) never had any trouble getting above the numbers I have been getting with my 965BE/R9-280 and Crysis 3 on high. I don't have the exact numbers right now, but I will try to get some soon.

@ Toyota
Were you on HIGH or on VERY HIGH? My system spec was HIGH. And SMAAx1

 

[toyota]

My other machine (i5-2500K/R9-290X) never had any trouble getting above the numbers I have been getting with my 965BE/R9-280 and Crysis 3 on high. I don't have the exact numbers right now, but I will try to get some soon.

@ Toyota
Were you on HIGH or on VERY HIGH? My system spec was HIGH. And SMAAx1

very high with medium shadows and SMAA x1

 

[AtenRa]

well I am looking at Crysis 3 now using up to 90% of my oced 4770k probably averaging 65%. it makes no sense that a Phenom 2 was not even getting maxed out and he was getting in low 40s for framerate. he must be testing at the least cpu bound part of the game. that game ate my oced 2500k for lunch at times and its was obvious.

EDIT: screen I took from last night


how to print screen on pc

I DO feel better about the card/CPU combo now. I kept MSI Afterburner on the other screen and monitored the GPU and CPU usage in percent, as well as framerate and frametime.

He used two Monitors, when he was monitoring the CPU and GPU utilization he was in full screen mode and not like you in window mode. There is a difference in performance and CPU and GPU Utilization between the two modes in the majority of games.

 

[toughtrasher]

He used two Monitors, when he was monitoring the CPU and GPU utilization he was in full screen mode and not like you in window mode. There is a difference in performance and CPU and GPU Utilization between the two modes in the majority of games.

Just noticed this.. Never actually looked at the stats from fullscreen to windowed.

 

[brokencage]

I have a phenom II B50 OC'd to 980 levels and I get a touch of bottle necking on a few games with a r9 270. oc'd @ 1225/1500. I wouldn't really go past a gtx 760 or a amd R9 270/270x.

 

[Clutchc]

OK. For comparison (and for toyota) I promised I would try to get some results using my other machine (2500K/290X). I couldn't get the 2nd monitor stay on to read Afterburner, so I had to resort to Fraps. I managed to run part of the same section that I ran for the Phenom II test. I tried to run for the same amount of time I was watching the graphs with MSI AB. These are the results from Fraps:

2014-05-18 12:34:00 - crysis3
Frames: 1370 - Time: 118499ms - Avg: 59.728 - Min: 50 - Max: 62

Appreciably faster that the 3.74GHZ PhIIX4/R9-280 was. As expected. In my limited experience, it all looks logical to me. I did the best I could to keep it apples to apples.

 

[toyota]

lol you must have found the least cpu intensive part of the game to have 50 fps min. when I first played Crysis 3 using the 2500k, I did not realize it was cpu intensive until I kept seeing gpu usage drop so much. on just high settings with medium shadows I could run the game at 1280x720 or 1920x1080 and it made little difference in many areas.

 

[Clutchc]

lol you must have found the least cpu intensive part of the game to have 50 fps min. when I first played Crysis 3 using the 2500k, I did not realize it was cpu intensive until I kept seeing gpu usage drop so much. on just high settings with medium shadows I could run the game at 1280x720 or 1920x1080 and it made little difference in many areas.

I don't know what to say. I played the entire game twice with that 15-2500K rig (once with the R9-290X and once before that with 2 x Evga GTX 660 SC cards) at the same settings. I never had any issues that hampered smooth game play. The only issue I had was occasionally a little stutter with the SLI setup. I blamed that on...SLI.