GTX 970 With Phenom II Not So Bad

[scirishman76]

Running A Phenom II 955 @3.8Ghz and just upgraded from a GTX 480 To a GTX 970 and did some test along the way. I tried to read up before hand but mixed bag of opinions some saying bad CPU bottleneck and others saying not really. Watched several video's of the Phenom II and 970 where GPU would not drop below 99% and none of the CPU cores were pinned at 90% or above so decided to give it a go.
I ran several test's both synthetics and in game fraps tests right before I put in the 970 and right After. I"ll list results in two columns one below 480 and one below 970. Hope this helps seemed to be alot of confusion when I was trying to read on it.


Heaven 4.0 1920x1200 dx 11 GTX480 score: 559 :: GTX 970 1192
ultra/extreme tess/8XAA min/max/av:480- 12/43/22:: 970- 19/99/46

Unigine Valley score: GTX480 1037 :: GTX 970 2001
min/max/av: 480-14/45/25 :: 970- 21/95/48

Firestrike Score: 480- 3183 :: 970- 8154

Farcry 4 min/max/av:480- 27/42/35:: 970- 57/104/78

Metro LL min/max/av:480- 28/45/36 :: 970- 47/96/80

BF4 Bench Av FPS 3 runs GTX480- 38-38-40:: GTX970- 69-69-71

Warthunder min/max/av:480- 32/69/46:: 970- 67/107/86

All Game tests were done @1920X1200 with settings that stressed my 480 left them alone for testing to get an accurate performance increase Keep in mind also my 480 had a huge OC and the 970 is stock.
Most tests I saw right around double the performance when switching to the GTX 970 I would say if your running a gtx 480/570/660 or AMD 6970/7850/R9 270 or 270X with a Phenom II don't be afraid of this upgrade. Also searched results databases in benches that allowed me and found that this upgrade scaled the same as it does with an I7 Haswell meaning that an I7 upgrading from a 480 to a 970 sees about the same 90% to 120% increase that my Phenom II did.

Sorry for the long post but when I searched for this seemed to be alot of confusion and no real results.

 

[Raiin]

go play AC, Watch dogs, a x4 955 does not scale the same as Any haswell CPU not even an i3, upgrading to Sandy,Ivy or haswell will yield some serious FPS increase.

BF4 is seriously low, you should easily be getting an average of at least 90-100FPS on ultra at 1200p

 

[scirishman76]

Raiin totally misunderstood scale. The the performance difference is similar % wise on these CPU's. Never argued that the I7 or I5 wouldn't outperform just argued that the upgrade was similar for both. Also if you can honesty show me a 320$ CPU and Mobo upgrade that would have given me 100% increase in FPS Im all ears, guessing Link saw about a 25-40% increase in Frames and paid more considering it requires a mobo change.
I know my CPU is dated but lots of reasons to wait right now with Intel just getting into DDR4 and AMD working on ZEN I feel if I upgraded CPU and Mobo not I would probably regret it within a 7-8 months. However I knew I wouldn't regret the GTX970 and it seems it should be able to keep me playing high to max settings until I see what unfolds for a CPU.
Also Raiin how can you say My BF4 frames are low when you don't know where I have my settings. You know turning the resolution scale way up will bring pretty much any GPU to under 50 frames. and Mine is turned up quite a bit.

 

[madmatt30]

If you're happy with performance that's all that matters IMO.
What you're getting is similar performance to a fx4300.
There's no doubt an i5 would pull 30 or so more fps & have less drops to minimum but who cares really?
You've bought a future proof card so the option is always there to upgrade the rest of the system in the future.I wouldn't consider it a waste of money at all.

 

[scirishman76]

I can shoot you numbers If you like based on an I7 4770 and a Phenom II x4 both upgrading from a 480 to a 970 in percentage here goes.
Firestrike I7 272% PII 267%
Unigine Heaven I7 109% PII 112%
BF4 I7 85% PII 82%
Stalker COP I7 94% PII 98%
Ran All the 3dmarks and had similar results to firestrike. And All the Unigine with similar results. Would list more but these were all I could find results databases for. Seems to scale about the same by my definition of scale.

 

[scirishman76]

Never intended to be an AMD vs intel thread why do people insist on making it that. Thats what the word SCALE that Link and Raiin don't seem to understand was meant to avoid this. Bottom line no 320$ upgrade in CPU and Mobo would have given me even half the gains in FPS. If you want to buy a new Mobo and CPU right now with all the changes that are coming in the next 7-12 months be my guest.

 

[scirishman76]

Never intended to be an AMD vs intel thread why do people insist on making it that. Thats what the word SCALE that Link and Raiin don't seem to understand was meant to avoid this. Bottom line no 320$ upgrade in CPU and Mobo would have given me even half the gains in FPS. If you want to buy a new Mobo and CPU right now with all the changes that are coming in the next 7-12 months be my guest.

Point is due to your CPU you could have saved money on GPU and got the exact same performance.

Not really in my case considering I am planning on upgrading everything as soon as Im comfortable that its a good choice. Even If I had bought a 4770K (almost did) would have still bought the 970.

 

[scirishman76]

No Use arguing Just showed that scale wise it was not a bottleneck and you can see for yourself video's all over the net that a 970 can pin at 99% usage on a Phenom II that never goes over 75% on any single core so by definition bottleneck is misused. The 4770K has an average of 25-30% increase over a phenom II across the board with some lopsided situations for both, Watchdogs and BF4 being great examples I7 has about 55-60% edge in watchdogs but in BF4 it is only 2-5% http://www.techspot.com/review/734-battlefield-4-benchmarks/page6.html http://www.techspot.com/review/827-watch-dogs-benchmarks/page5.html

Maybe this will help clear things up I7 4770 is the same 25-30% better pushing a gtx 460 as it is pushing a GTX 970 not a bottleneck just a cpu that pushes 25-30% harder in most games. Unless you are now gonna claim that a Phenom II Bottlenecks a GTX 460.

 

[scirishman76]

Also worth noting that my coworker and I had almost identical systems he had a Phenom II and a GTX 570 my Phenom II GTX 480 setup. We ran almost Identical bench scores and FPS in every game. He Upgraded to A 4770K and a fairly high end Mobo about 5 months ago his cost $569. I just spent $320 and have 50-70% lead in FPS and bench scores that we have compared. So while I agree that of you are in a position to upgrade both by all means do. However I don't see how spending more money for less gains can be justified.
Link Your being able to buy a cheaper GPU is really pushing it. Kinda sounds like lets spend $500 on a cpu mobo combo so I can spend $250 instead of $320 to get the same performance as the guy that just spent the $320 and saved the $500 lol.

 

[AlmightySofa]

The people in these forums astound me with their stubbornness sometimes, they keep throwing bottleneck around like it is holy water. What does it take to get you guys to understand that it is NOT a bottleneck unless the CPU load is 100% while the GPU load is significantly less. Otherwise the performance difference is simply a battle between which cpu is faster. If you pay more, you get more fps, that is all. By your definitions of bottleneck, if a 4460 is slower than a 4690k then the 4460 is a bottleneck... no? The 4690k is just a bit faster? People these days.

 

[scirishman76]

Link says cpu for performance gpu for quality hmmmm wrong again. For gaming gpu holds more weight period. Go buy a high end cpu and a crappy gpu and your games will run crappy. Buy a weak cpu and a great gpu and it will work alright.
As far as image quality its pretty standard info that two Nvidia cards running the same settings will look identical even if one is fairly low end and onethe other is really high end. Same goes for AMD gpus image quality controlled by settings.

 

[madmatt30]

The latest 2 pc game releases are being stunted by directx 11 api's draw call limitations more than any hardware.
Ravens cry & dying light both pull 50-60% CPU usage on my fx6300 & 970 GPU never gets above 50% use in game .both games force big draw calls & directx is unable to cope.
I myself am running a 970 on a 6300@4.2ghz & I'm more than happy with performance.
I vsync everything to 60fps anyway, & my 970 stays under 80% usage & below 60c this way.
Said before anyone running without vsync on a 60htz screen & forcing 100fps+ ,100% GPU usage & high temps for no reason apart from bragging is an idiot anyway.

 

[AlmightySofa]

Link what is your point? It's common intel is faster per clock but it doesn't change the fact that the amd cpu isn't a bottleneck.

 

[scirishman76]

Getting Dying Light soon Ill post performance here as soon as I get it.

 

[scirishman76]

Link says cpu for performance gpu for quality hmmmm wrong again. For gaming gpu holds more weight period. Go buy a high end cpu and a crappy gpu and your games will run crappy.

That's why when I had the Phenom II proc and the same HD 7850 as I do today with Intel Core i5 ... with just the Intel CPU swap I can play any game @60fps min ... with AMD Phenom II I was often stuck at well below 60 fps no matter what settings used because of CPU bottleneck.

True, However you could have sunk far less money into a new GPU and seen far greater performance gains as far as FPS so how again does CPU equal more performance than GPU which seems to be what your saying.
Bottom line you spent 350-450 on a cpu mobo and gained probably 20-30% in FPS I spent 320$ and saw a 110% gain. Don't see how you can even argue.

 

[scirishman76]

Link what is your point? It's common intel is faster per clock but it doesn't change the fact that the amd cpu isn't a bottleneck.

I owned the exact CPU as OP right before Christmas before upgrading to Intel. You are going to sit here and tell me Phenom II is not a bottleneck LOL.

I have defined Bottleneck a few ways and Link still doesn't get it. Lets try this By definition A cpu Bottleneck means you would see little to no gain from upgrading other hardware cause your CPU simply can't feed the parts so they cannot perform at a 100% usage. My GPU does perform at 100% usage and I saw a 110% gain how is that a bottleneck.

 

[scirishman76]

Link what is your point? It's common intel is faster per clock but it doesn't change the fact that the amd cpu isn't a bottleneck.

I owned the exact CPU as OP right before Christmas before upgrading to Intel. You are going to sit here and tell me Phenom II is not a bottleneck LOL.

I have defined Bottleneck a few ways and Link still doesn't get it. Lets try this By definition A cpu Bottleneck means you would see little to no gain from upgrading other hardware cause your CPU simply can't feed the parts so they cannot perform at a 100% usage. My GPU does perform at 100% usage and I saw a 110% gain how is that a bottleneck.

Sigh. I will let a moderator explain it to you then.

Been Building/tuning working on PC's and in the PC industry since before some of the moderators in here were born. I don't think I need someone to explain a bottleneck to me.

 

[scirishman76]

Didn't bother you came into my thread and tried telling me that a CPU upgrade that would have yielded me 15-35% increase in FPS for more money was a better upgrade path than a GPU upgrade that costs less and yielded me over a 100% fps increase.
Your right games run great on Intel. However If upgraded my CPU I would choose the I7 4770/4790K combined with my old GTX 480. This setup would be equal or greater to your I5 and 7850 setup for sure and to be brutally honest has no chance of keeping up with my Phenom II with a GTX 970 in games. You spent more money for about 1/3 the performance increase and somehow feel that you can justify it by jumping all over me in my thread. I made this thread to show folks with Phenom II CPU's not to be afraid of upgrading to the GTX 970 not to try to start an AMD v Intel war.

Just Tested Dying Light 1920X1200 Everything Maxxed
Tower Min/Max/av 39/133/67Outside Min/Max/Av 37/127/62

 

[madmatt30]

^ you won't average those kind of results once you've played a while - there's no consistency at points within the game - I've seen 17fps once with barely anything happening. Other times it sits at 60fps solid with like 20 zombies on screen.
Poorly optimised bag o sh1te ,good game though & good looking.
More interested in you CPU/GPU usage in game tbh.
On mine even when it hits lows fps wise the usage on either very rarely gets above 60%.

 

[scirishman76]

^ you won't average those kind of results once you've played a while - there's no consistency at points within the game - I've seen 17fps once with barely anything happening. Other times it sits at 60fps solid with like 20 zombies on screen.
Poorly optimised bag o sh1te ,good game though & good looking.
More interested in you CPU/GPU usage in game tbh.
On mine even when it hits lows fps wise the usage on either very rarely gets above 60%.

Those results were averaged from about 12 tests over 4 and 1/2 hours of gameplay have been playing a while so I dont expect to see the low FPS dips you are speaking of. And My GPU pinned at 98-99% except for a couple of real short dips to 75ish guessing for a load screen or something. Maybe the two are connected and you are spiking low because your GPU isn't ramping up.

 

[madmatt30]

^ interesting ,what os are you running??
I'm on 8.1
There are others with the same issue as me though & others who aren't.
This issue is only on dying light ,even with vsync off I'm only hitting 60% GPU usage ,the CPU is steadily between 50-60% on all cores.
There's no throttling ,its just not utilising the GPU properly.
I can max GPU usage out on any other game but generally run 60htz vsynced as its pointless not doing on a 60 htz screen.
I normally hit 60fps vsync on other stuff like farcry 4/advanced warfare etc with 70% GPU usage & 60c temps which is why I bought the card.

I'll try the game on my 8320/r9 280x windows 7 setup at the weekend to see if this is the same.
The games perfectly playable & the GPU sits at like 50c while playing but the fps drops for no apparent reason are annoying .

 

[scirishman76]

Using Win 7 64 will test CPU usage tonight after work for ya.

 

[scirishman76]

Link you will be hard pressed to find a modern game that is so cpu dependent that it would allow a gtx 480 or hd 7850 to come close to keeping up with a gtx 970. I agree to some extent though I do want to upgrade my cpu mobo, but as I said earlier I like to make my upgrades count and beacause of that I feel like waiting 6 months to 1 year for intel ddr4 or AMD Zen to develope more is the best choice.

 

[Mousemonkey]

Hmm, I wonder what a 970 and an E8400 might run like?

 

[AlmightySofa]

Don't bother with this Link guy. Nobody was saying anything about the Phenom cpu the OP had. I was stating that fact that you cannot determine that it is a cpu bottleneck if the GPU is not working at 100% percent. Period. End of story.