• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

AMD FX 8320 analysis

We don't know if we need an i7 for SLI, it wouldn't do sweet anything over an i5 if the game can't use the extra threads.
And as mentioned it's using an effect that kills the frame rate.

And I can't imagine multi-gpu's working great tbh.

I disagree from my experience. 2 power gpu's or more, impose 20-25% more load on the cpu load with vsyn off on an intel 4 core cpu in BF4.

Heres a point in case 3750k 4.6Ghz, cpu usage circa 75% 1x7950 1200Mhz 99-100% gpu usage with vsyn off. With vsyn on (60hz) still circa 70-75%...(cpu not bottlenecked)

Now add another card

2x7950's 1200Mhz, cpu usage jumps from 75% with vsyn off to 95-100%.
Turn vsyn (60Hz) on guess what happens to cpu usage it is pegged back at 75-80%.

Also if I turn cards down with vsyn off cpu usage returns to circa 90%.

So point in case proven

2xHD 7000+ or 680's will bottleneck an i5 with vsyn off. (Not with vsyn on 60hz)
 
Last edited:
I don't think you understand.

BF4's a completely different engine, and it's proven to be threaded and be able to use 8 cores.

All an i7 has over an i5 in 4 threaded (We don't know how many threads, or how the engine works, we can't say an i7 will/will not be required) is 2mb cache. BF4's a completely different scenario with different factors.
 
Unless you are referring to specific scaling in this game only. I though your statement was generic about 4 cores and 4 cores with threading. What is the case you were making?

For that game specifically ; COH2.

I know quite clearly about bottlenecks, I'm pretty sure I explained to you about the i5 when you were adamant it wasn't a bottleneck.
 
Last edited:
For that game specifically ; COH2.

I know quite clearly about bottlenecks, I'm pretty sure I explained to you about the i5 when you were adamant it wasn't a bottleneck.

A blind statement saying an i5 is a bottleneck in all situation with sli/crossfire doe not really show what happens when uses operate 60hz monitors operating at 60fps.

How powerful the gpu's are is a very important factor as to is screen refresh rate's another is in game settings another is whether the game scales efficiently(as you mention above).

Some people who play games like bf4 at 40fps want to know if they can use a second gpu to attain 60fps or more and if they believe the mass hype then the answer is that an i5 is a total bottleneck. The modern era i5's do not bottleneck 2xgpu's with vsyn on fullstop. That is it full stop. If your running powerful gpu's at 120hz or with vsyn off then its an i7 period....
 
I never give a blanket statement, in BF3/BF4 the i5's a bottleneck with multi GPU's like 7950's like you were using, that's what I meant, as that was the exchange we had.

Running Vsync is just hiding the bottleneck.
 
Last edited:
I never give a blanket statement, in BF3/BF4 the i5's a bottleneck with multi GPU's like 7950's like you were using, that's what I meant, as that was the exchange we had.

Running Vsync is just hiding the bottleneck.

Vsyn does remove bottleneck, no one is hiding the potential for a bottleneck.

It is possible for 2 relatively powerful gpu's to cause a brief cpu bottleneck if it hits 100% with vsyn off, but can you truely regard 120-140fps with vsyn off as a major bottleneck for a 3570k. The capability of my monitor is 60hz so having 2.5x that refresh rate is hardly a huge issue as my cards still seem to max out and if anything the gpu's reach their thermal limits with such a large oc, whereas the cpu keeps going hour after hour.

Having Vsyn off in my case is:

a) a total waste of energy in terms of power consumption due to overclock voltage.

b) unperceptable or marginally perceptable at best with my 60hz refresh rate.

My statement remains as above.
 
In lots of games you gain nothing at all with an i7 over an i5 with two GPUs, e.g. here

http://images.anandtech.com/graphs/graph6934/54523.png

Multithreaded games that are quite CPU intensive (e.g. BF3 multiplayer) are different. If a game only uses 1-2 cores there's unlikely to be any difference. It's got nothing to do with vsync.

Oh yeah, I agree.
There's more games that the i5 will be fine multi-gpu than there isn't.

Using Vsync doesn't remove the bottleneck, it masks/hides it. The bottleneck is still there, but you won't see it as you're preventing it, but that's semantics, you see it as bottleneck gone, I see it as bottleneck hidden.
 
Last edited:
In lots of games you gain nothing at all with an i7 over an i5 with two GPUs, e.g. here

http://images.anandtech.com/graphs/graph6934/54523.png

Multithreaded games that are quite CPU intensive (e.g. BF3 multiplayer) are different. If a game only uses 1-2 cores there's unlikely to be any difference. It's got nothing to do with vsync.

Sleeping Dogs dont scale so this is a game that is poorly optimised for multigpu usage, so your right no benefit here. If a game uses 1-2 cores but is gpu limited then a more powerful gpu is required. If the game is cpu limited due to poor cpu utilisation then your pretty much stuck becuase its a code programming restriction.

Battlefield 4 and Rome 2 scale with additional gpu's. Infact BF4 has a 200fps maximum and people routinely hit this using 2 or more gpu's.

I've hit a max of 160fps with 2x7950's with an ivy i5.
 
Last edited:
Rome 2 is a mess, it's bottlenecked by the CPU drastically.

Sometimes I just think people don't understand bottlenecks etc.

Yeah Rome 2 annoys me too, but it still scales well with additonal gpu's run the bench mark to confirm this. (Although the benchmark fps performance shows no resemblance to in game performance)
 
So Rome 2 doesn't really scale well, because it's so CPU bottlenecked, apart from the benchmark, which means nothing (It just puts the emphasis on the GPU from what I can tell)

Games don't tend to scale all that bad with Multi-GPU, the only reason they would is a CPU limitation of crap drivers really.
 
So Rome 2 doesn't really scale well, because it's so CPU bottlenecked, apart from the benchmark, which means nothing (It just puts the emphasis on the GPU from what I can tell)

Games don't tend to scale all that bad with Multi-GPU, the only reason they would is a CPU limitation of crap drivers really.

Only when you zoom in on large scale battles is there a major issue. Try the benchmark as a pointer. Your right the game has issues but the benchmark shows how the game might run when its fixed.
Anyway did you buy the blood pack? Its terrible...lol
 
I'd hope the game runs better than the benchmark as a single 7970 is only getting like 30 FPS.

I played it on launch, it's far better than it was, it's actually playable on maximum settings with a 7970 now.
I haven't bought the blood pack as I'm waiting for them to fix the game.
This is why I never buy games on launch, but I broke my rules for Football Manager 2014 and Rome 2, regret them both.
 
I'd hope the game runs better than the benchmark as a single 7970 is only getting like 30 FPS.

I played it on launch, it's far better than it was, it's actually playable on maximum settings with a 7970 now.
I haven't bought the blood pack as I'm waiting for them to fix the game.
This is why I never buy games on launch, but I broke my rules for Football Manager 2014 and Rome 2, regret them both.

Yeah Rome 2 is the biggest let down in my gaming history. Seige forked, no family tree, AI dumb, no seasons. Its a disaster. People on Totalwar forum just constantly slate CA and Sega and rightly so IMO. Apparently they are not going to fix the Seige AI and wont introduce family tree or seasons.

I have had it with TW.
 
Sega have been ruining CA for a while, and I know the game would be crap.
Low and behold, Sega all behind Football Manager 2014.

EA are ruining MMO's and FPS etc.
Sega are running games that need any thought process.

And Ubisoft are just hashing the same old formula.
 
Back
Top Bottom