AMD Fury X Far Cry 4 game performance from AMD

[Opcode]

The benchmarks on the Vessel video today show the 980 Ti Stock getting ~10 More FPS than was shown in the Benchmarks in the OP...

It's a G1, factory overclocked GTX 980 Ti that's why it's beating even the TITAN X. I presume OC would mean switched into OC mode.

 

GTX 980 Ti G1

 

Base

• 1190MHz (OC Mode)
• 1152MHz (Gaming Mode)

Boost

• 1291MHz (OC Mode)
• 1241MHz (Gaming Mode)

[niofalpha]

 

It's a G1, factory overclocked GTX 980 Ti that's why it's beating even the TITAN X. I presume OC would mean switched into OC mode.

 

GTX 980 Ti G1

 

Base

• 1190MHz (OC Mode)
• 1152MHz (Gaming Mode)

Boost

• 1291MHz (OC Mode)
• 1241MHz (Gaming Mode)

 

Yes, the stock 980 Ti Still gets 10 more frames in Far Cry 4 than in the benchmarks shown in OP tho.

[Opcode]

Yes, the stock 980 Ti Still gets 10 more frames in Far Cry 4 than in the benchmarks shown in OP tho.

That's the problem when comparing results from different sources. There's a lot of variables to take into consideration such as other hardware, drivers, etc. We won't really know how Fiji performs until someone sits down and benches Fury X, TITAN X, GTX 980 Ti and a handful of other cards in the same system with all of the latest WHQL released drivers. Although even still Fury X is out performing a G1 GTX 980 Ti which is still pretty impressive at much lower clocks. The biggest problem is there's only one source and one result for Fury X performance and that being AMD. It's too bad they won't just lift NDA already and let the gates flood with benchmarks.

[niofalpha]

That's the problem when comparing results from different sources. There's a lot of variables to take into consideration such as other hardware, drivers, etc. We won't really know how Fiji performs until someone sits down and benches Fury X, TITAN X, GTX 980 Ti and a handful of other cards in the same system with all of the latest WHQL released drivers. Although even still Fury X is out performing a G1 GTX 980 Ti which is still pretty impressive at much lower clocks.

We haven't seen the Benchmark AMD ran on the Fury X either. For all we know it could just be staring at the sky. 

[Jerakl]

I'm gonna go with those benchmarks are a load of bullshit on this one.

 

I wonder how much power it's gonna draw, I'm guessing that you're gonna need at least a 1000W PSU to have two in crossfire.

 

 

--Note to those who don't understand things--

The whole 1000W PSU for two fury xs in CF was a jab at amd as their cards tend to be on the higher power consumption side of things.

 

...

Is an example of a completely unnecessary comment.

Nazi mod alert!

[zMeul]

That is what's expected with GDDR5. HBM and AMD's implementation will help solve that.

interesting, who told you that?!  :blink:

[Diceair]

Impressive. Getting that fps on Far cry 4. Unoptimized game and even dx11 and not even overclocked. Guess when overclocked this thing will be a beast. You all must remember the Memory will also make contact with the AIO cooler so you can overclock both. I guess AMD might win this time and I might just get this card. The only thing that might keep me from buying is the whole thing about ghosting on freesync monitor. Although cheaper than g-sync it's still expensive and I'm scared about that issue.

 

If AMD can pull that off and we can get a card with DVI included they have instant buyer but if there won't be a DVI version then sorry. I run a Qnix and must have active displayport to DVI cable (one with usb power) and then I can't overclock it so it's a real deal breaker. If only the panel makers can sort out the ghosting it's instant buy for me. 

[zMeul]

That is what's expected with GDDR5. HBM and AMD's implementation will help solve that.

no it won't solve s**t because system RAM will still be DDR3 in most cases, and DDR4 for the few

as I said countless times already, HBM is only relevant in GPU<->VRAM communication

[MegaDave91]

 

It's a G1, factory overclocked GTX 980 Ti that's why it's beating even the TITAN X. I presume OC would mean switched into OC mode.

 

GTX 980 Ti G1

 

Base

• 1190MHz (OC Mode)
• 1152MHz (Gaming Mode)

Boost

• 1291MHz (OC Mode)
• 1241MHz (Gaming Mode)

 

I'll reiterate this again. LTT's Titan X is a terrible sample which boosts 12mhz higher than the guaranteed boost clock. Watch this video where Jay puts the Titan X and stock G1 980ti at an even (basically even) 1400mhz. In fact, his results show the Titan X beating he OC G1 980ti as well.

 

[FakezZ]

Am I the only one that noticed the asterisk in the "Ultra Settings" in the picture? What could that mean?

[Opcode]

I'll reiterate this again. LTT's Titan X is a terrible sample which boosts 12mhz higher than the guaranteed boost clock. Watch this video where Jay puts the Titan X and stock G1 980ti at even (basically even) 1400mhz. In fact, his results show the Titan X beating he OC G1 980ti as well.

 

You cannot reiterate something that hasn't been said. Jays numbers only validate what I've been saying. There's too many variables surrounding just one number without any direct comparison provided within the same source. These numbers can give us a peek into what Fiji is capable of but we won't really know until independent reviews. According to Jays numbers both GM200 cards are still -14 FPS slower than Fury X in FC4 at 4k.

 

Am I the only one that noticed the asterisk in the "Ultra Settings" in the picture? What could that mean?

The bottom of the slide explains, we just need to dig up the last few slides wherever they are...

 

[Dabombinable]

no it won't solve s**t because system RAM will still be DDR3 in most cases, and DDR4 for the few

as I said countless times already, HBM is only relevant in GPU<->VRAM communication

Your being an idiot. I was talking in regards to vRAM (which I've been talking about all along, with you going off half cock about 'pop in' when playing The Witcher 3. When its not related at all to the BS your spouting about PCIe Gen 2.0 (it is not a bottleneck for any graphics card)-because I've said it before and I'll say it again. Pop in is caused by a bug or glitch in the game engine. And TW3 has already had loads of patches).

Plus your perpetuating the myth that 4GB of vRAM isn't enough. Even though the current performance benchmarks of the Fury X with HBM state otherwise. (That and the benchmarks with SLI GTX 980 at 4K vs the Titan X. 4GB vRAM vs 12GB vRAM).

[zMeul]

...

PCIe bandwidth wasn't even a point because, also, system ram is still the same

do we have HBM system RAM and I missed it?!?! then ... what are we talking about

funny how @RononDex understood perfectly what I was trying to say 

[Opcode]

PCIe bandwidth wasn't even a point because, also, system ram is still the same

do we have HBM system RAM and I missed it?!?! then ... what are we talking about

funny how @RononDex understood perfectly what I was trying to say 

Lets hear your explanation in detail as to how system memory is a limiting factor to Fiji performance now.

[MegaDave91]

You cannot reiterate something that hasn't been said. Jays numbers only validate what I've been saying. There's too many variables surrounding just one number without any direct comparison provided within the same source. These numbers can give us a peek into what Fiji is capable of but we won't really know until independent reviews. According to Jays numbers both GM200 cards are still -14 FPS slower than Fury X in FC4 at 4k.

 

 

 

I'm re-iterating what I've said. I'm not comparing either to the Fury X, rather the Titan X to the G1 980ti.

 

As for the Fury X, I won't dwell too much on the numbers AMD has presented. Let's wait until independent reviewers get their hands on it.

[RononDex]

funny how @RononDex understood perfectly what I was trying to say 

 

zMeul at some point it makes no sense to have a discussion / argument anymore. Dabobminable seems to get pretty agressive / insulting.

I haven't been following the whole conversation but when someone starts calling you "idiot" and such I would immediatly stop a discussion as the partner is definitly not mature enough to have a mature discussion.

This has nothing to do with the topic you are actually talking about, it's a common thing that people like to get personal / insulting when they feel they are beeing put into a corner and can't put anything against someone's argument. If this happens do yourself a favor and stop the discussion.

[Opcode]

I'm re-iterating what I've said. I'm not comparing either to the Fury X, rather the Titan X to the G1 980ti.

First time quoting (replying to) me in this thread, I wouldn't call that reiterating. Why bother compare the GTX 980 Ti G1 and the TITAN X. The idea is to dig numbers for both cards and compared them against AMD's projected numbers. I don't think anyone in this particular thread cares about Nvidia vs Nvidia.

[MegaDave91]

First time quoting (replying to) me in this thread, 

That must mean I've said it elsewhere. 

[zMeul]

zMeul at some point it makes no sense to have a discussion / argument anymore. Dabobminable seems to get pretty agressive / insulting.

I haven't been following the whole conversation but when someone starts calling you "idiot" and such I would immediatly stop a discussion as the partner is definitly not mature enough to have a mature discussion.

This has nothing to do with the topic you are actually talking about, it's a common thing that people like to get personal / insulting when they feel they are beeing put into a corner and can't put anything against someone's argument. If this happens do yourself a favor and stop the discussion.

so true ...

but lately, that ignore list has gone up in numbers, quite fast since yesterday  

[MegaDave91]

so true ...

but lately, that ignore list has gone up in numbers, quite fast since yesterday  

That's to a tee why I generally lurk these types threads instead of contribute. I don't have anyone on my ignore list, though. 

[Dabombinable]

PCIe bandwidth wasn't even a point because, also, system ram is still the same

do we have HBM system RAM and I missed it?!?! then ... what are we talking about

funny how @RononDex understood perfectly what I was trying to say 

Do I have to drag up all of your posts in regards to PCIe bandwidth? Ok then.....

 

naming and shaming .. how cute

but you failed my furry friend because all that CG users showed in the AMD sub-forum is that they lack understanding of basic computer hardware architecture

not even going further into matters where they compare multi-GPU with single GPU in some idiotic comparison, bluntly ignoring multi GPU systems work in alternate frame rendering

 

good job there buddy! good job

here's a cookie: https://upload.wikimedia.org/wikipedia/commons/6/6e/Pepperidge-Farm-Nantucket-Cookie.jpg

now, get back to CG forums and tell them how you wiped the floor with me - not! but maybe they'll give you a cookie too, just for effort

 

maybe you'll have time to compare diplomas, since you know .. e-peen measuring

 

if I'd known in advance I'd have this discussion, I'd be prepared with recorded footage

but demand me to pop with evidence just like snapping fingers .. wow!

 

no one even bothered testing for this and bam! ... I'm instantly wrong  :blink:

 

yes it is because unified RAM architecture gets rid of the need to make system RAM - VRAM swaps

thus, you have very concrete DRR3 vs GDDR5 comparison

 

if you fail to understand that ... not my problem

Not a source I like to use-but plenty of references: https://en.wikipedia.org/wiki/List_of_device_bit_rates

 

when you go from 16x to 8x you don't see any drop? oh yes you do - even when going from PCIe gen3 to gen2

You being more of an idiot:

 

there is but one tiny issue with that theory

HBM only improves bandwidth between GPU and VRAM, not between system and video card

did the PCIe changed? no! did system RAM changed? no! so, in the grand scheme of the system as a whole, HBM is rather irrelevant

a total game changer would be when we'll have HBM far system RAM and a CPU to support it

---

are those benchmarks true? probably are

but there should be benchmarks with games that truly have hi-rez assets for UDH resolutions, and up

 

the problem is not the difference is little, the problem is that there is a difference

and as you go up in resolution that PCIe bandwidth matters even more

just compare TW3 on PCIe gen3 vs gen2 and see how objects will have pop-in on gen2 - extremely noticeable

---

here is one more example on why HBM is kinda irrelevant:

XB1 vs PS4 - XB1 uses DDR3 while PS4 uses GDDR5, both as unified RAM

and yet, there is very little difference between them

Proof you ignored....

 

There's very little difference regardless of the slot you place a card (1-3 FPS). Linus has done at least three videos about this.

 

 

 

 

 

FYI, that's 15 pages I've gone through.

 

[Homicidium]

That's to a tee why I generally lurk these types threads instead of contribute. I don't have anyone on my ignore list, though. 

 

Me neither, because I have the same approach to most of these topics, as you do. It's in everyone's best interest to stay the f out threads like these, because EVERY DAMN TIME some people need to derail these type of threads with their fanboy-ish and immature behaviour.  :rolleyes:

[Opcode]

That must mean I've said it elsewhere. 

Not that I've seen, nothing worth reiterating in the first place in an AMD oriented thread. We only care about Fury X performance against the competition, not so much TITAN X vs 980 Ti. Whatever results we come by some having both GM200 products benchmarked in them, we just go by whichever performs better. So far it's been an onslaught with Nvidia unable to match said performance until you start heavily overclocking their products. Although we're still talking about only one game, one result and without any direct comparisons to other products. So everything is still what I like to call, inconclusive.

 

Although I don't mind liking to think that Fiji is just going to pulverize GM200, we'll find out soon enough.

 

so true ...

but lately, that ignore list has gone up in numbers, quite fast since yesterday  

Ignore lists do you no favors if you actually think about it. If you ignore someone they can still quote you and pick at your posts meanwhile you remain oblivious to it.

[Dabombinable]

zMeul at some point it makes no sense to have a discussion / argument anymore. Dabobminable seems to get pretty agressive / insulting.

I haven't been following the whole conversation but when someone starts calling you "idiot" and such I would immediatly stop a discussion as the partner is definitly not mature enough to have a mature discussion.

This has nothing to do with the topic you are actually talking about, it's a common thing that people like to get personal / insulting when they feel they are beeing put into a corner and can't put anything against someone's argument. If this happens do yourself a favor and stop the discussion.

I say things as they are. When some one is being an idiot I tell them so, and provide proof to go with it. Also, read my above post and the very same style of comment from @zMeul that you are accusing me of making.

[shawno]

Id like to hear @patrickjp93 s comments on the pci/ram/hbm debacle and what sort of limitations there would be.