AMD A10 5800K+HD6670 (dual Graphic) vs HD7790

[yanuar rizky pramusesa]

Sorry for my english..i just want to ask, which one better for my A10 5800K? for the lattest game and graphic design..should i use dual graphic or stand alone VGA?thanks

 

[atomicWAR]

7790 will be better than hybrid crossfire by a very good chunk.

 

[stickmansam]

7790 wins out

CFX only scales from 50-80% so 150-180%
The 7790 is 2.5x faster than a single 7790

 

[atomicWAR]

can't argue with benchmarks...nice addition to the argument stickman.

 

[stickmansam]

gotta love graphs 😛

 

[atomicWAR]

and you 😉 jk i just think your a good tech is all

 

[blazorthon]

No offense, but the benchmark and comment didn't make sense when I actually read them. I assume that stickmansam meant to say that the 7790 is 2.5 times faster than a single 6670 and that was the mistake which would most certainly prove the point about the 7790 being faster, but it still invalidates what was said about CFX.

Yes, the Radeon 7790 would be a much higher performance graphics solution.

Crossfire's scaling has far to many factors to make any generalization at all without much more specific details. However, the 7790 would still win because even with perfect scaling between the most highly overclocked A10-5800K's GPU with the fastest memory and the fastest Radeon 6670 with the fastest overclock on GPU and memory, the Radeon 7790 at stock would still be faster in almost all, if not all, gaming situations and probably almost all, if not all, other situations too.

Crossfire does not only scale 50-80% nor does CrossfireX (the distinction being that CFX is the three/four GPU extension of Crossfire, not a different name for the same exact thing). They are both capable of having much worse and somewhat better scaling. I've seen anywhere between 50% and even down to negative scaling just as at least with CF, I've seen ~100% scaling at times. There are many factors involved such as architecture of the GPU, umber of GPUs, drivers, driver profiles, the game, the game's version, and so much more.

 

[atomicWAR]

No offense, but the benchmark and comment didn't make sense when I actually read them. I assume that stickmansam meant to say that the 7790 is 2.5 times faster than a single 6670 and that was the mistake which would most certainly prove the point about the 7790 being faster, but it still invalidates what was said about CFX.

Yes, the Radeon 7790 would be a much higher performance graphics solution.

Crossfire's scaling has far to many factors to make any generalization at all without much more specific details. However, the 7790 would still win because even with perfect scaling between the most highly overclocked A10-5800K's GPU with the fastest memory and the fastest Radeon 6670 with the fastest overclock on GPU and memory, the Radeon 7790 at stock would still be faster in almost all, if not all, gaming situations and probably almost all, if not all, other situations too.

Crossfire does not only scale 50-80% nor does CrossfireX (the distinction being that CFX is the three/four GPU extension of Crossfire, not a different name for the same exact thing). They are both capable of having much worse and somewhat better scaling. I've seen anywhere between 50% and even down to negative scaling just as at least with CF, I've seen ~100% scaling at times. There are many factors involved such as architecture of the GPU, umber of GPUs, drivers, driver profiles, the game, the game's version, and so much more.

makes perfect sense....not sure what your missing. it shows 6670 is not up to par with 7790 and he states the scaling issues that the graph doesn't cover. i prefer sli if going multi-gpu. less buggy and usually better scaling.

 

[stickmansam]

I was making the assumption that the game was a scalable onethat sees benefits and the most I've seen is 50-80% (very rarely you'll actually get 100% scaling)

 

[blazorthon]

No offense, but the benchmark and comment didn't make sense when I actually read them. I assume that stickmansam meant to say that the 7790 is 2.5 times faster than a single 6670 and that was the mistake which would most certainly prove the point about the 7790 being faster, but it still invalidates what was said about CFX.

Yes, the Radeon 7790 would be a much higher performance graphics solution.

Crossfire's scaling has far to many factors to make any generalization at all without much more specific details. However, the 7790 would still win because even with perfect scaling between the most highly overclocked A10-5800K's GPU with the fastest memory and the fastest Radeon 6670 with the fastest overclock on GPU and memory, the Radeon 7790 at stock would still be faster in almost all, if not all, gaming situations and probably almost all, if not all, other situations too.

Crossfire does not only scale 50-80% nor does CrossfireX (the distinction being that CFX is the three/four GPU extension of Crossfire, not a different name for the same exact thing). They are both capable of having much worse and somewhat better scaling. I've seen anywhere between 50% and even down to negative scaling just as at least with CF, I've seen ~100% scaling at times. There are many factors involved such as architecture of the GPU, umber of GPUs, drivers, driver profiles, the game, the game's version, and so much more.

makes perfect sense....not sure what your missing. it shows 6670 is not up to par with 7790 and he states the scaling issues that the graph doesn't cover. i prefer sli if going multi-gpu. less buggy and usually better scaling.

[citation] The 7790 is 2.5x faster than a single 7790 [/citation]

[citation] CFX [/citation]

[citation] CFX only scales from 50-80%b [/citation]

[citation] CFX only scales from 50-80% so 150-180%[/citation]

No, it didn't make sense. The first citation states that the 7790 is faster than a single 7790, an obvious typo, but still wrong nonetheless. The second is wrong because CFX isn't in use here, that would be CF or more accurately, because a Trinity APU is in use, it's just called Dual Graphics. The thirs part specifically states that CFX only scales from 50-80%. There are several things wrong with that, but I'll just tackle the main issue in that CF nor CFX nor Dual Graphics nor any other such form of multi-GPU technology from Ati and AMD only scales within that range.

The last is wrong in my opinion rather than fact in that it gives the implication on the graph that we should look in the 150% to 180% faster than the base of 100% despite the fact that we should be looking for something 150% to 180% faster than the 6670 itself. Also worth mentioning is that the 6670 in that graph is a GDDR5 version despite the fact that we only need to care about DDR3 versions for Dual-graphics because the GDDR5 version is less suited for it.

Another thing that I find to be worth mentioning is that none of this was pointed out, so OP may have assumed that the data in that graph was relevant for comparing the performance that OP would get out of a Dual Graphics solution with the A10-5800K and a Radeon 6670. That's taking shortcuts in giving advice and I really try to avoid someone not getting the fullest idea that I can give, within reason, when I'm giving advice.

None of that is meant to be offensive to anyone and I apologize if it does offend anyone. Even if it's selfish of me, I wasn't going to just deal with being told something that is nonsensical does make sense and I'm deficient in some way for not thinking so. I may be over-thinking it at this point, but I'm still technically correct.

One last set of points to make: SLI does not usually scale better than Crossfire, at least not with modern cards (modern being VLIW4 or GCN GPU-based cards from AMD and Fermi or Kepler cards from Nvidia, especially with Fermi) and even more important is that both this and especially how buggy each is depends on driver version, game, and more, so no wide generalization is important nor even likely to matter.

 

[blazorthon]

I was making the assumption that the game was a scalable onethat sees benefits and the most I've seen is 50-80% (very rarely you'll actually get 100% scaling)

I don't have a problem with that. I only have a problem with what you said in that because you said it only scales 50-80% rather than it usually scales 50-80% or something with similar meaning, especially since you didn't specify something like Crossfire with those two GPUs and current drivers or such because Crossfire's scaling tendencies can vary greatly between different GPUs, different GPU architecture, different drivers, and that's just a start without even going into games (which can also be impacted in scaling by settings and more).

I'm not sayign that it's bad to not overload the OP with info, just that a little info would be more helpful IMO and even if not giving more info, not saying something incorrect like only in that context would be helpful too IMO.

 

[stickmansam]

I need to be more clear in my posts then I guess 😛

The point was to drive home to the OP though was that a 6670+a10 5800k would be decisively beaten by a 7790

 

[mohit9206]

how about A10+6670 vs 7750 ?

 

[stickmansam]

I would get an a8 5600k + 7750 over an a10 + 6670

at best the dual graphics can only match the 7750 but most of the time the 7750 will be faster. How much faster depends on the specific gamr and settings.

the a10+ 6670 is cheaper than an a10/a8 + 7750

 

[twelve25]

plain and simple, dual graphics crossfire with an APU is a waste of time. It's a gimmick. It's barely worth it even if you already own a compatible card.

Try and find some benchmarks that show dual graphics with a 6670 even beating just a 6670 alone. I can't.

 

[mohit9206]

plain and simple, dual graphics crossfire with an APU is a waste of time. It's a gimmick. It's barely worth it even if you already own a compatible card.

Try and find some benchmarks that show dual graphics with a 6670 even beating just a 6670 alone. I can't.
you are kidding right ? games that support crosssfire show a very good scaling with dual graphics. not 100% scaling but nearly double the frame rate of the APU itself.

 

[twelve25]

double the frame rate of the APU, but so does a standalone 6670. The point being, using dual graphics offers no benefit over just using a single standalone card.

as an example, 1fps less than the A10 with just the 6670

This one actually shows dual graphics doing well in bf3 vs an i7 and a 6670, but then in the second game on that page it does MUCH worse: http://www.rage3d.com/reviews/fusion/amd_a10_5800k_launch_review/index.php?p=11

So really it's not something I would ever recommend.

 

[mohit9206]

so if a person had to choose a dual core pentium +7750 or just an A10 for gaming what should he choose ? on one hand the dual core pentium would soon be outdated since it has only 2 cores , and on the other the A10 has a weak gpu. personally i was in the same situation and i chose the first option as you can see in my sig. however this also means i will have to upgrade my cpu sooner which will cost extra.

 

[twelve25]

If you are spending under $150 on graphics + processor it is out of date instantly, regardless. Some games will perform well with a limited processor, but NO games perform well with a crummy video card. I'd go dual core + graphics card every time.

 

[stickmansam]

I would rather run a a10/a8 for a while and then add a graphics chip later

The pentiums are going to be dated real soon

 

[darkspartenwarrior]

In some cases Hybrid Crossfire does absolutely nothing for framerate, only adding 1-2 FPS.

The 7790 is a much better choice, smoother and faster.

 

[twelve25]

I would rather run a a10/a8 for a while and then add a graphics chip later

The pentiums are going to be dated real soon

Again, if you are JUST running an APU, then you have no dream of playing games like crysis 3 at high framerates anyway. I'd like to see dual core with a decent GPU vs the APU alone. That chart paired it with a 7970, didn't it?


What about the other 90% of games where the graphics card is the bottleneck?

And you could just as easily buy an i5 as an upgrade to a pentium as you can add a GPU to an A10. If we are assuming the person will have money for upgrades later, then I wouldn't want to stick them with a low end FM2 socket. Then they top out at A10/FX-4300/i3 level instead of having access to I7/i5 or FX-8300 series.

 

[stickmansam]

I would rather run a a10/a8 for a while and then add a graphics chip later

The pentiums are going to be dated real soon

Again, if you are JUST running an APU, then you have no dream of playing games like crysis 3 at high framerates anyway. I'd like to see dual core with a decent GPU vs the APU alone. That chart paired it with a 7970, didn't it?


What about the other 90% of games where the graphics card is the bottleneck?

And you could just as easily buy an i5 as an upgrade to a pentium as you can add a GPU to an A10. If we are assuming the person will have money for upgrades later, then I wouldn't want to stick them with a low end FM2 socket. Then they top out at A10/FX-4300/i3 level instead of having access to I7/i5 or FX-8300 series.

Again the same budget issue. If the OPhas the budget to get an i7/i5 + GPU then by all means

What I meant is the OP is on a tight budget, going an A10 now, then getting an stronger GPU is a better way to go than an pentium + 6670. Then upgrading the GPU and CPU which will cost more.

 

[twelve25]

Of course there are different ways of looking at it. I can see the point you are making in regards to upgrade path. But, what about performance on day 1? I just don't see starting out with a rig that already *requires* an upgrade to be a decent gamer. I'd rather have issues in a handful of games that like a quad core than issues in EVERY game.

 

[mohit9206]

The question is if you have only $150 to spend for both gpu and cpu which is a better long term solution for a gamer ?
a pentium cpu + radeon 7750 or an A10 or A8+6670 ? (assume 720p fixed resolution)