New 6970, but am I hardware limited?

[apoppin]

Really? I thought I saw benchmarks to the contrary. Something like SB having double the FPS of Penryn. Perhaps that was at low res, I don't recall.

If that's true, then it's good to know. I built my BIL a computer a year or two ago, with a Q9550 @ 3.5, and a pair of HD6870 cards on an X48 mobo. Was wondering if I should consider upgrading his computer, or whether it will be good enough for future games for the next couple of years.

My personal opinion is that the rig is pretty damn powerful as-is, although slightly less powerful than an i5-2500K @ 4.4-4.5 and a 6990.

An i5-2500K at 4.5GHz is mostly overkill for a 6990 above 1920x1080.

Any benches that you see run with a big disparity, *always* have the games benched at 1680x1050 (or below) with no AA and (mostly) medium details.

For gamers, there is no practical difference from running a Penryn Quad or Phenom II X-4 to a Sandy Bridge setup - as long as the CPUs are reasonably (overclocked) and the resolution and detail settings are maxed out; And i have tested this with GTX 580 SLI and HD 6990 Tri-Fire using my i7-920 from 2.6 to 3.6GHz compared with Phenom II and Penryn.
^_^

 

[jamesbond007]

Of course i did; you are getting all excited about your Penryn Quad core somehow being inferior to SB (Sandy Bridge, not SoB ) in one game and moaning that you don't have enough money to upgrade it.

I wasn't saying you didn't read my comments - just the GrumpyMan poster! And I said that money wasn't a problem, but justifying the cost to just one game is not worth it to me.

Sandy Bridge makes no practical difference in gaming over a fast Penryn Quad with a decent graphics card at 1920x1200. Those are the facts. i don't know about your specific case.

I wouldn't think so, but clock for clock, the SB is stronger. All I can tell you so far with my 15 minutes of 'testing' last night (I said I had my GF over!) is that by increasing my OC on my CPU, I get more performance and removing my OC, I lose FPS. That is a direct relationship and I can adjust my graphics quality (AA/AF) and I see no difference in performance in WoW.

That said, I will run 3DMark2011 tonight so I can give you guys some real numbers to see where everything is at. Would a dxdiag output be helpful to you? I will run the 3DMark11 on both the 'test' OS install HDD and my working one as well.

Thanks!

 

[apoppin]

Not just 3DMark11 .. try a few benchmarks - and ideally some other (non-online) games you want to play.

Yes, clock-for-clock, Sandy Bridge is stronger than Penryn. But not so much so to make a 2X difference in frame rates at 19x12.

Perhaps you might be able to increase your current OC by getting a better CPU cooler; i had my old Q9550S up to 4.0Ghz. 3.4 to 3.6GHz seemed to be the sweet spot for a HD 6970 class of card.

 

[RussianSensation]

Sandy Bridge makes no practical difference in gaming over a fast Penryn Quad with a decent graphics card at 1920x1200. Those are the facts.

False. Depends on the game. My stock i7 860 was 40-45% faster in Resident Evil 5 at 1920x1080 8AA with a lowly HD4890 compared to my Q6600 @ 3.4ghz (not a Penryn but performance is very comparable).

It simply depends on the game. While I agree with you that 2-3x difference can't come from a CPU alone (esp. when discussing a Penryn @ 3.2ghz), but there is definitely a difference in games between my i7 860 @ 3.9ghz and the Q6600 @ 3.4ghz I had. Therefore, there should undoubtedly be a difference between a 2500k @ 4.5ghz and a Penryn at 3.2ghz, even at 1920x1200. This will be especially true if you are running an Nvidia videocard.

Core i5 750 vs. i7 975 shows a difference, nevermind a SB vs. Penryn.

 

[apoppin]

False. Depends on the game. My stock i7 860 was 40-45% faster in Resident Evil 5 at 1920x1080 8AA with a lowly HD4890 compared to my Q6600 @ 3.4ghz (not a Penryn but performance is very comparable).

It simply depends on the game. While I agree with you that 2-3x difference can't come from a CPU alone (esp. when discussing a Penryn @ 3.2ghz), but there is definitely a difference in games between my i7 860 @ 3.9ghz and the Q6600 @ 3.4ghz I had. Therefore, there should undoubtedly be a difference between a 2500k @ 4.5ghz and a Penryn at 3.2ghz, even at 1920x1200. This will be especially true if you are running an Nvidia videocard.

Core i5 750 vs. i7 975 shows a difference, nevermind a SB vs. Penryn.

And how was that 40% difference in RE5 practical? It runs pretty well on a Q6600/HD 4890. And you not only changed platforms (not even from a Penryn to an i5), you upped the CPU clockspeed by 500MHz.

The issue is PRACTICAL differences in gaming ... and from your own link:

The CPU does not affect the frame rate much in this game and its influence lowers at higher resolutions. The Core i7-975 EE is 6-9% ahead of the Core i5-750 in terms of average speed at 1600x900, but at 2560x1600 the gap is only 4% with the dual-processor card and within 1% with the single-GPU Radeon HD 5000 series models. It is at the lowest resolutions that the CPU affects the bottom speed the most. For the Radeon HD 5770, replacing the CPU with a faster one can already make sense then, but we guess that purchasing a faster graphics card, e.g. a Radeon HD 5850, will be an even better solution.

There just isn't that much practical difference in gaming between Penryn and SB (at higher detail/high resolutions); and as you admitted, nothing like the 2x-3x difference the OP is reporting.

 

[jamesbond007]

Yes, clock-for-clock, Sandy Bridge is stronger than Penryn. But not so much so to make a 2X difference in frame rates at 19x12.

Perhaps you might be able to increase your current OC by getting a better CPU cooler; i had my old Q9550S up to 4.0Ghz. 3.4 to 3.6GHz seemed to be the sweet spot for a HD 6970 class of card.

You would be surprised at how poorly coded the WoW engine has been reported to be. Please keep in mind that it is over 6 years old now and I'm quite excited for the 64bit client to be released, but I'm not expecting too much from that. As far as the CPU cooler goes, I'm already on a semi-water cooling with 35C being the highest I have ever seen it with Prime95, so I'm pretty sure that's non-issue.

 

[apoppin]

Not that poorly coded. People play it on their notebooks. There can be latency issues also. And there will never be 2x to 3x performance difference running a Sandy Bridge CPU at a similar clockspeed to a Penryn Quad.

However, if you are seeing 35C as your CPU's highest temp, then push your OC higher. And give us your other benchmark results, please.

 

[jamesbond007]

Not that poorly coded. People play it on their notebooks. There can be latency issues also. And there will never be 2x to 3x performance difference running a Sandy Bridge CPU at a similar clockspeed to a Penryn Quad.

However, if you are seeing 35C as your CPU's highest temp, then push your OC higher. And give us your other benchmark results, please.

Yeah I play it on my MacBook, barely. But hey, it DOES work, but boy is it bad. I would LOVE to give you numbers, but I'm working late!

I should mention that now my lowest FPS is about 30, my friend's machine would be about 2x faster.

 

[MangoX]

The improvement not only comes from a faster cpu, but from having a better chipset, moving from slower DDR2 to faster DDR3 memory , and from FSB to QuickPath etc...

 

[apoppin]

The improvement not only comes from a faster cpu, but from having a better chipset, moving from slower DDR2 to faster DDR3 memory , and from FSB to QuickPath etc...

Basically agreed although x48 uses DDR3 and Penryn is pretty advanced architecture with very little advantages for SB in gaming. Even so, there should only be a few percentage difference at the same clockspeets - not 200-300% difference in FPS as the OP is noting.

Anandtech reviewed WoW and there isn't much difference like the OP is seeing.
http://www.anandtech.com/show/4083/...core-i7-2600k-i5-2500k-core-i3-2100-tested/20

We run on a Radeon HD 5870 at 1680 x 1050. We're using WoW's high quality defaults but with weather intensity turned down all the way.

 

[peonyu]

All,

I got a pretty good deal on a 6970 2GB that I couldn't pass up to swap out my heaterblock in my computer, AKA Radeon 4870 512MB. After doing a driver cleaning, it does not seem to be much faster than my 4870 in some regards.

I play at 1920x1200 in games, but I could never crank up the AA or AF because the 4870 would choke after 2x at that resolution. With the new card, I can put it on any setting and it does not lower my FPS at all...

I don't have any benchmarks yet from 3DMark or anything else, but I do use WoW to bench. I recently built my friend a new computer with an Intel i7-2600K CPU, 16GB RAM, same Radeon 6970 2GB, and the ASUS Sabertooth P67 motherboard. His FPS is nearly double or triple of what I get at the same settings! I have tried changing back and forth from DX9 and DX11, but no dice.

In the main cities I will get between 25-30fps with every possible setting maxed, sometimes dipping into the high teens, which isn't right. Last night I hooked up a spare hard drive, installed Windows 7 fresh and downloaded WoW, but I got the exact same results. (yes, I did turn off all my addons and I never installed any on my temporary Windows install last night) I even tried his video card, but I yielded the same numbers.

The only thing I can conclude is that I seem to be CPU limited some how or perhaps there is a bandwidth limitation of the X48 chipset I have. We have traveled to a secluded area and his FPS will still be almost always double that of mine. I can run some other benches and other games later if you'd like, but I just can't fathom the fact I would be limited in some CPU-related way at 1920x1200, even though my processor doesn't go above 25-35% CPU while playing.

My computer components:
Intel Core 2 Quad 9450 2.66GHz @ 3.2GHz
Corsair H50 'Water-cooling' (temps are ~35C at load)
ASUS Rampage Formula motherboard
8GB G.Skill 800MHz @ PC1066MHz
Sapphire Radeon 6970 2GB @ stock
OCZ Vertex 2 SSD
PC Power and Cooling 750W Silencer

I did remove my overclock last night, but that lowered my FPS more, which makes me believe even more that I am somehow limited by my CPU or motherboard.

Thanks in advance!

WoW can run great on a 5 yr old card, its all cpu limited today. Crank up the AA and enjoy.

 

[MangoX]

Basically agreed although x48 uses DDR3 and Penryn is pretty advanced architecture with very little advantages for SB in gaming. Even so, there should only be a few percentage difference at the same clockspeets - not 200-300% difference in FPS as the OP is noting.

Anandtech reviewed WoW and there isn't much difference like the OP is seeing.
http://www.anandtech.com/show/4083/...core-i7-2600k-i5-2500k-core-i3-2100-tested/20

I know you review hardware and all for a living, but I gotta ask... have you ever played WoW? I played WoW from vanilla all the way until Catalysm launched. There were too many changes in that expansion that I decided I didn't want to spend so much more time re-learning everything and quit with my Draenai Warrior (main) at lvl 80. I have witnessed performance improvements thru every CPU upgrade.

 

[apoppin]

I know you review hardware and all for a living, but I gotta ask... have you ever played WoW? I played WoW from vanilla all the way until Catalysm launched. There were too many changes in that expansion that I decided I didn't want to spend so much more time re-learning everything and quit with my Draenai Warrior (main) at lvl 80. I have witnessed performance improvements thru every CPU upgrade.

Yes i have. About 2 years ago. i really didn't like it.

What's your point? i have a lot of difficulty believing that a Sandy Bridge Quad CPU would make a 2-3X performance improvement in framerates over a Penryn Quad CPU - everything else being the same.
:\

 

[MangoX]

Yes i have. About 2 years ago. i really didn't like it.

What's your point? i have a lot of difficulty believing that a Sandy Bridge Quad CPU would make a 2-3X performance improvement in framerates over a Penryn Quad CPU - everything else being the same.
:\

In the Anandtech review you linked there isn't a q9450 in there. I know the q9450 is about 10-15% faster than the q6600. Based on the other benches, the X4 645 appears comparable to the q6600, so we'll add 13% on top the X4 645 to simulate the q9450.

On the WoW bench, the X4 is getting 62.9fps while the 2600K is getting 119.4fps. Simulating a q9450 (62.9*1.13) we come up with 71 fps. That's a 168% increase.

But in that benchmark they're only testing in an open, empty area with no other players. In a 25-man raid during a trash mob/boss fight, I wouldn't be surprised to see a 2X possible increase.

 

[apoppin]

In the Anandtech review you linked, the q6600 appears comparable to the X4 645, and the WoW graph (it's missing the q6600), the X4 is getting 62.9fps while the 2600K is getting 119.4fps. To me that's a near 2X boost.

But in that benchmark they're only testing in an open, empty area with no other players. In a 25-man raid during a trash mob/boss fight, I wouldn't be surprised to see a bigger boost larger than 2X.

What are you talking about?

The OP has a Penryn Q9450 Quad-core clocked at 3.2 to 3.6GHz. His friend has a similar Sandy Bridge CPU system with the same video card (HD 6970) that he claims is 2x to 3x faster only because of the CPU.
:\

And he has done no further benchmarking so we don't know if there are other issues or not including networking. It is pretty hard to draw any conclusions at this point from a single online game.

 

[MangoX]

I edited my post when I realized the q6600 is a Kentsfield core >_<

 

[apoppin]

I edited my post when I realized the q6600 is a Kentsfield core >_<

OK. i'd suspect architecture could favor Sandy Bridge over Kentsfield or Phenom II in specific games. However, finding major performance differences between Penryn and Sandy Bridge at similar clockspeeds is quite a stretch in gaming.

It is possible that WoW is an exception. i don't play the game.

 

[MangoX]

It is hard to prove. Not many sites bench the game and even then, they aren't in a raid scenario which is probably the most taxing event in the game. I'm just trying to share my experiences being a ex-hardcore WoW player and relating to the OP. And it's hard to do that when there isn't any reliable data to show.

 

[jamesbond007]

So I finally ran 3DMark2011! I received "P5074 3DMarks". This was with the basic edition running the Performance preset and "Benchmarks only" selected. My results can be viewed here.

Any thoughts?

 

[RussianSensation]

So I finally ran 3DMark2011! I received "P5074 3DMarks". This was with the basic edition running the Performance preset and "Benchmarks only" selected. My results can be viewed here.

Any thoughts?

A stock 6970 should get around 5400-5500 points. So you are not that far off. 10% difference compared to a Core i7 system. So definitely your videocard is NOT the issue. Have you tested any other games? WOW is just 1 game, with horrible graphics to begin with btw. Based on the graphics quality (welcome to 2004), you should be getting 60+ fps with your videocard. In fact, an HD6970 gets 76 fps at 1920x1080 4AA/16AF. If that's the only game that runs slow on your system, I wouldn't worry about it (unless this is the only game you play - in which case I feel bad for you!). Make sure the Power is +20% in your catalyst control panel, so that your videocard is not throttling back. Also, check that your CPU is not thermal throttling either. What are your CPU/GPU temperatures and the voltage on 12V rail when the GPU is loaded?