AMD CPU Gaming benchmarks

[SiliconWars]

I just wonder why AMD and Intel release those sub $200 CPUs ?? Everyone could spend $100 or more if they could wait a coudple of weeks to add to the pot

Sub $200 CPU's probably outsell $250-$300 CPU's by a factor of 5 to 1 actually, which is why AMD is still very relevant.

 

[AtenRa]

Customer : Hello, I would like a gaming PC

PC shop : What is your budget ??

Customer : $600.00

PC Shop : Could you come back in a few weeks with $750 to suggest you a gaming PC with a faster CPU ??

LoL

 

[boxleitnerb]

Well, then there is something absolutely wrong with benchmarks. I just tested both CPU's with the same OCed 7950 and found the 8350 to be faster in f.i TR And my intension was to give it a fair go with either CPU to see what CPU to use for games.I've spend hours and hours of testing and I played TR with both CPU's and found no difference in gameplay. Anyway...it seems benchmark results are dependent of lot more than just CPU's? And only owners of both CPU's, AMD and Intel, can test and test and find out.
Don't forget Anandtech and other sites use the best motherboards/setups available but for consumers this is a bit different. Consumers test them in a more practical manner with different hardware.

You mean Tomb Raider? Tomb Raider is fairly GPU bottlenecked, especially with SSAA and/or TressFX. Btw if you're using vsync, it's a moot point anyway. No difference in gameplay would point to a GPU bottleneck, thus equal results.

The difference between those CPUs becomes apparent in other games. There are titles where it really doesn't matter which CPU you use, and I believe TR is one of them.

 

[cmdrdredd]

I just wonder why AMD and Intel release those sub $200 CPUs ?? Everyone could spend $100 or more if they could wait a coudple of weeks to add to the pot

Cause people buying those CPUs generally aren't concerned with them being ultra high performance for gaming applications and other potentially CPU intensive work.

If you are spending a lot of money already, someone suggesting an extra $100 for a better part isn't that far fetched. If you don't feel you have the budget that's fine, but the suggestion is given with no bad intentions attached.

 

[PPB]

Lol cinebench as a metric to show gaming performance. It cant even show a good picture of 3D production rendering performance, let alone gaming.

 

[Makaveli]

Yes I did first time around. That said, the 955 would also be pushed to near 4.2 GHz meaning there wouldn't be that much difference while overclocked. It's not like you get linear gains from oc'ing a CPU in gaming benchmarks.

I think you are underestimating how much of a performance gain you will get with the extra 1200-1400 mhz on the intel chip compared to the 500+600mhz gain you would on the 955.

I would like to see more numbers but I think the gap will be larger with both chips at 4ghz or 4.2.

 

[SiliconWars]

I think you are underestimating how much of a performance gain you will get with the extra 1200-1400 mhz on the intel chip compared to the 500+600mhz gain you would on the 955.

I would like to see more numbers but I think the gap will be larger with both chips at 4ghz or 4.2.

I was being generous to the i5 actually. That i5 760 will be running at 3.3 GHz turbo for most of those benchmarks so overclocking to 3.8 GHz will only give a small 15% benefit in single threaded games. The 955 would be overclocked by 30% at 4.2 GHz.

Note I'm taking the numbers from this post - http://forums.anandtech.com/showpost.php?p=35261476&postcount=49, assuming 3.8 GHz for the Lynnfield and 4.2 GHz for the Deneb.

 

[toyota]

Well, we should start making CPU gaming performance conclusions having a Cinebench 11.5 benchmark run from now on.




Well, i have 600.00 euros to spend not 730.00, ill take an FX6300 + HD7950 than Core i5 + HD7790. For two-three years i will enjoy my games at higher image quality than you and after that when i will really need a new GPU i will consider my options at that time. You on the other hand will be all this time with inferior image quality but dont worry, after two-three years you will be able to game at the same settings as i did all those years before you when you will finelly upgrade the GPU.

I believe you got my point, but im sure you will continue tooting about the faster and more expensive CPU.

so you will "enjoy" playing many games at well under 60, 50, and even 40 fps with a 7950. so your logic is that because you cant afford to buy balanced system that you should spend more on the gpu and bottleneck the crap out of it from day 1. brilliant!

if building a gaming pc from scratch that will use higher end cards then the only sensible choice is a 4670 or 4770. I do enjoy the laughable excuses and scenarios people come up with to buy an inferior cpu that will already limit current gpus nevermind the next gpu that is upgraded too.

 

[wilds]

I'm willing to wager that a 8320 and 8350 are good purchases. The problem is that many games are not using 8 threads.

After looking at Crysis 3, and Metro: Last Light, the 8350 matches the 4770K very closely.

Those two games are very CPU heavy.

Buying an Intel i5/i7 will guarantee you high performance while buying an FX CPU seems to be at the game engine's mercy.

 

[Blitzvogel]

OC

http://pclab.pl/art51730-14.html
http://pclab.pl/art53333-14.html

looking at the core i3 is easy to conclude the game is not using many cores,

I don't think FSX is a well coded multicore game. From my experience and what I've seen/heard, it's more like the game just runs almost all it's code on a single core with maybe some utilization of a second.

I've ran the game across a few systems:

Turion x2 1.8 GHz + Geforce 7200M
Athlon x2 5600 3.0 GHz + Geforce 8800GTS 320 MB
Core 2 Duo P8600 2.26 GHz + Geforce 9800M GS 512 MB
Phenom II x4 925 2.8 GHz + Radeon 5850 1 GB [Current system]
Core i3-330M + Geforce 310M 256 MB

Regardless of improvement in CPU architectures and core counts, it ran like ass, even with the graphics and environment/world settings set way low. Never understood why MS allowed for such a sloppy piece of programming. I'd much rather just stick with FS2004 if I'm gonna play a FS game.

 

[SPBHM]

After looking at Crysis 3, and Metro: Last Light, the 8350 matches the 4770K very closely.

Those two games are very CPU heavy.

4770K = 930?

edit:

and

 

[SiliconWars]

The problem with cherry picking benchmarks is there are always others that show you cherry picked.

Note I just googled "metro last light gaming benchmarks" and "Crysis 3 cpu benchmarks" and techspot was the first site on each, so I went with that.

 

[SPBHM]

Note I just googled "metro last light gaming benchmarks" and "Crysis 3 cpu benchmarks" and techspot was the first site on each, so I went with that.

the test I posted from Crysis 3 is newer, and possibly using a newer version (which fixed the low performance with the i7 in "welcome to the jungle") yours 1.0, also with higher settings and faster graphics card, it also specifies the scene and it was made in 2 different scenes, it also includes the 4770k and proves the 4770k and 8350 are not comparable for crysis 3, which was his claim.

if we accept that both tests are valid, but the one I posted is more relevant (newer, includes Haswell, more scenes), we still have a clear loser here...

 

[wilds]

the test I posted from Crysis 3 is newer, and possibly using a newer version (which fixed the low performance with the i7 in "welcome to the jungle") yours 1.0, also with higher settings and faster graphics card, it also specifies the scene and it was made in 2 different scenes, it also includes the 4770k and proves the 4770k and 8350 are not comparable for crysis 3, which was his claim.

if we accept that both tests are valid, but the one I posted is more relevant (newer, includes Haswell, more scenes), we still have a clear loser here...

The 8350 costs considerably less and still delivers relatively good performance in games. It comes close to the 4770k while using more power, is less efficient per watt, and has awful single threaded performance.

I would easily choose an Intel component. But if I was budget constrained, I would gladly choose the 8350 because it is sufficient for games and I believe it will be able to run games fluidly for a long time.

In fact, I think the Piledriver is a relatively good architecture. I notice the quad-core variant performs relatively well as well.

 

[SPBHM]

The 8350 costs considerably less and still delivers relatively good performance in games. It comes close to the 4770k while using more power, is less efficient per watt, and has awful single threaded performance.

I would easily choose an Intel component. But if I was budget constrained, I would gladly choose the 8350 because it is sufficient for games and I believe it will be able to run games fluidly for a long time.

In fact, I think the Piledriver is a relatively good architecture. I notice the quad-core variant performs relatively well as well.

it doesn't cost considerably less than core i5s, and it's clearly slower for most games (page 1, 2 on this topic is a good example), comparable for others (like on Crysis 3 for the most part but in one scene you can see the 5GHz 8350 behind the stock 3570)... I didn't choose the comparison with the 4770k (which is more expensive) I was just answering this: "the 8350 matches the 4770K very closely."

the 8350 looks good for the money for some other work (like 3d rendering) but when it comes to gaming that's not the case, you can think it's good enough, but that doesn't mean it's faster, better or as good.

as I said before, the 6300 offers more for the money than the i3 3220 or whatever costs the same... just like the FM2 Athlon X4 does... but when you start comparing Piledriver to Core i5 for gaming it's a different case.

 

[escrow4]

Prices in Australia:

8350 = $229 + $121 for a decent 970 board with phases - M5A97 EVO R2 = $350

4670K = $269 + $145 for Asrock Z87 Pro 4 = $414

Difference is $64, why bother at that difference, the i5 sucks way less power and is faster in modern games.

That 6300 is $135 here, the closest Intel would be the the 4430 @ $209. At $269 vs $135 I can see why you'd choose AMD and sacrifice a lot of frames, but only if you are really really broke.

I also built myself a basic office box with a G1610 - an A4 4000 Richland build would have cost me $20 more and sucked way more power while being substantially slower than the Intel. There isn't that much value in AMD over here at least for the 8350.

 

[monstercameron]

Prices in Australia:

8350 = $229 + $121 for a decent 970 board with phases - M5A97 EVO R2 = $350

4670K = $269 + $145 for Asrock Z87 Pro 4 = $414

Difference is $64, why bother at that difference, the i5 sucks way less power and is faster in modern games.

That 6300 is $135 here, the closest Intel would be the the 4430 @ $209. At $269 vs $135 I can see why you'd choose AMD and sacrifice a lot of frames, but only if you are really really broke.

I also built myself a basic office box with a G1610 - an A4 4000 Richland build would have cost me $20 more and sucked way more power while being substantially slower than the Intel. There isn't that much value in AMD over here at least for the 8350.

questionable...

 

[justin4pack]

what i am seeing is that the fx-6300 comes pretty close to 8350 performance if overclocked slightly. its also $70 cheaper.

 

[Hypertag]

Because some people just dont have it. I should have guessed this would turn into an intel/amd debate. Listen compairing an fx6300 to a high end intel is sensless when i am compairing ONLY amd for.price performance. It needs be cheap but still perform good.

Whatever price penalty supposedly exists on Intel parts, most of it is just an illusion.

You can get a quad core ivy bridge processor with a decent motherboard for $251 if you just stop paying for the ability to overclock

Core i5 3330 $175 EMCXNXL35
http://www.newegg.com/Product/Produc...82E16819116781

Motherboard $76
http://www.newegg.com/Product/Produc...82E16813157303

The "massive" intel price penalty only exists if you pay $250 for a 4670k and $250 for an ultra high-end Z87 motherboard. It almost entirely ceases to exist if you compare the parts more accurately.

 

[Hypertag]

Customer : Hello, I would like a gaming PC

PC shop : What is your budget ??

Customer : $600.00

PC Shop : Could you come back in a few weeks with $750 to suggest you a gaming PC with a faster CPU ??

LoL

Yeah, actually yes. This is not Best Buy. Best Buy still exists, and serves its function. Here, we are building our own systems. Life won't end if you end up waiting two weeks for another paycheck in order to have enough money to properly build a system instead of taking shortcuts that are guaranteed and proven to cause further issues down the line.

 

[monstercameron]

Yeah, actually yes. This is not Best Buy. Best Buy still exists, and serves its function. Here, we are building our own systems. Life won't end if you end up waiting two weeks for another paycheck in order to have enough money to properly build a system instead of taking shortcuts that are guaranteed and proven to cause further issues down the line.

also questionable...

 

[cmdrdredd]

Whatever price penalty supposedly exists on Intel parts, most of it is just an illusion.

You can get a quad core ivy bridge processor with a decent motherboard for $251 if you just stop paying for the ability to overclock

Core i5 3330 $175 EMCXNXL35
http://www.newegg.com/Product/Produc...82E16819116781

Motherboard $76
http://www.newegg.com/Product/Produc...82E16813157303

The "massive" intel price penalty only exists if you pay $250 for a 4670k and $250 for an ultra high-end Z87 motherboard. It almost entirely ceases to exist if you compare the parts more accurately.

How is it accurate when people overclock AMD parts too? Might as well just compare the Intel part you will overclock.

also questionable...

You're also adding nothing to the conversation here. The post about prices in Australia may not be relevant in the states but it makes a point. The difference is $65 and the performance difference is definitely worth that. You just respond with "questionable..." which tells us nothing about your opinion and only serves to thread crap.

There was nothing questionable about the price comparison for AUS. Nothing at all.

 

[Obsoleet]

OP- The truth is, any new mid to top range chip from AMD or Intel released in 2013 is going to be fast enough and you likely won't see the difference. I use a CPU from 5 years ago. It's still going strong.

Both AMD and Intel CPUs are faster than what I have.
The market has stagnated, the idea you need a top end CPU to play games is ridiculous. If you benchmark more than you play games, by all means- sell the farm for a top end i7. I used to upgrade as often as I could afford to.

Buy whatever fits your price- not shoot for a narrow range of performance.
I game everyday, more than most of the CPU fanboys that post here I'm sure.

When I replace my Q9450, you'll know it's time to upgrade from one. Unless mine dies, I'm not switching it out till I really need to.

The game has changed, and the CPU forum on a hardware site is the last place you'll see that change being accepted.

I have a household income of well above $100,000 a year. I'm not the poor guy complaining about upgrading. I'm the guy who isn't an idiot building new rigs for no reasonable gain.

 

[monstercameron]

How is it accurate when people overclock AMD parts too? Might as well just compare the Intel part you will overclock.



You're also adding nothing to the conversation here. The post about prices in Australia may not be relevant in the states but it makes a point. The difference is $65 and the performance difference is definitely worth that. You just respond with "questionable..." which tells us nothing about your opinion and only serves to thread crap.

There was nothing questionable about the price comparison for AUS. Nothing at all.

look at the bolded parts in my prior posts.

Originally Posted by Hypertag
Yeah, actually yes. This is not Best Buy. Best Buy still exists, and serves its function. Here, we are building our own systems. Life won't end if you end up waiting two weeks for another paycheck in order to have enough money to properly build a system instead of taking shortcuts that are guaranteed and proven to cause further issues down the line.

that statement is questionable, if you haven't been reading the comments of the others, pretty much either way you go, you will be good given a powerful enough gpu, so if gaming is the criteria, a higher tier gpu should be sought after even to the potential detriment to single threaded cpu performance.

there you have it, my opinion, which has been stated many times in the thread alone by others.

 

[cmdrdredd]

look at the bolded parts in my prior posts.

that statement is questionable, if you haven't been reading the comments of the others, pretty much either way you go, you will be good given a powerful enough gpu, so if gaming is the criteria, a higher tier gpu should be sought after even to the potential detriment to single threaded cpu performance.

there you have it, my opinion, which has been stated many times in the thread alone by others.

And yet it took two thread crap posts before you actually said anything relevant.