No it was because Fermi was broken and delayed 6 months. Ati had the market to themselves.
-
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.
** The AMD Navi Thread **
- Thread starter subbytna
- Start date
No it was because Fermi was broken and delayed 6 months. Ati had the market to themselves.
HD 5870 wasn't that much faster than GTX 285, in the first place. The price was due to avoiding the internal competition with HD 4890, but I think the overall situation largely hurt AMD because people felt HD 5870 was overpriced.
Epeen, nothing more.
Soldato
- Joined
- 28 May 2007
- Posts
- 10,061
Yea overpriced at £300
. Now there is arguements whether a £1200 2080ti is overpriced. We have came a long way since then.If I remember rightly it was about 15% faster than a 285, 20% faster than a 4890, and 30% faster than a 4870.
285's were around £300 at the 5870 release, think the 4850x2 was around £250. Either way at £300 yes criticism was at the 5870 price but only because the 4870/4850 were so cheap on launch. Like I said Amd held the performance crown, were well practiced on the 40nm, and Nvidia fell on their arse. Amd priced the 5870 fairly it's just people wanted it far cheaper.
Soldato
- Joined
- 28 May 2007
- Posts
- 10,061
Was more like 30% faster than the gtx285. No less than 20% in some games and around 40% in others.
https://www.anandtech.com/show/2841/17
Yeah I was glad at the time I'd spent my money on a pair of GTX260-216 in SLI as they generally matched or beat the 5870 (SLI support was generally reasonable back then especially if you tweaked profiles, etc.) and IIRC cost me less than a GTX285 or maybe a little more I can't remember now.
Tbh I'm not worried about prices so much, because ultimately Nvidia can't strangle the GPU market so long as AMD's still fighting at all, and that's not going to change any time in the next 10 years, regardless of whether or not Nvidia has a GPU above AMD's best. The reason for that is simple: Nvidia still needs to sell GPUs, and their shareholders will want them to sell ever more. They can't increase prices willy-nilly without demand plummeting, that's a very simple business arithmetic, and indeed we haven't seen overall prices significantly increase for the performance levels below the highest end. That ultra enthusiast buyers are willing to part with £1.2k for +30% performance increase & the promise of RTX is not a big deal. The performance ceiling will always be infinite & unreachable, and therefore anything that inches you ever closer to it will command a premium. As true for GPUs as for anything else.
Let's also remember that there's a performance floor, namely consoles. And in about a year's time that floor will be established at 4K 60 fps for ~£500. 4K is already going into diminishing returns in terms of visual fidelity, certainly for mass market. So you can see that actually Nvidia is in a much more precarious position than their market dominant position would paint. No wonder then that they're pivoting so desperately into RTX. Ultimately that doesn't matter though, 4K60 will get even cheaper, RT or no RT, because it will be many years until games become exclusively RT'ed & until then raster performance will remain king. Though we might not even see non-RT capable GPUs (i.e. without at least some RT cores) anymore from next year except maybe at £150 and below.
All in all, I wouldn't worry. Nvidia can't keep up this pricing forever & AMD isn't that far away in terms of raw performance anyway.
Let's also remember that there's a performance floor, namely consoles. And in about a year's time that floor will be established at 4K 60 fps for ~£500. 4K is already going into diminishing returns in terms of visual fidelity, certainly for mass market. So you can see that actually Nvidia is in a much more precarious position than their market dominant position would paint. No wonder then that they're pivoting so desperately into RTX. Ultimately that doesn't matter though, 4K60 will get even cheaper, RT or no RT, because it will be many years until games become exclusively RT'ed & until then raster performance will remain king. Though we might not even see non-RT capable GPUs (i.e. without at least some RT cores) anymore from next year except maybe at £150 and below.
All in all, I wouldn't worry. Nvidia can't keep up this pricing forever & AMD isn't that far away in terms of raw performance anyway.
Associate
- Joined
- 17 Sep 2018
- Posts
- 1,431
Do we realistically seeing consoles doing this within 2 years though? Realistically that's at least 2080/Radeon 7 levels in a console. If that's the case we get Navi 20 GPUs in 2020 at £350 hitting 2080/Radeon 7 performance.
Unless it isn't real and they're going to employ some form of upscaling or DLSS
It is a pipe dream, unless like you say AMD will implement some form of DLSS on consoles.
The general point that the existence of console and simple market dynamics that will limit Nvidia's prices is true though. So i don;t see Nvidia's prices going too much higher. The market will also self-correct regardless of AMD in the PC space. If the market fails to support Nvidia's price point then Nvidia will simply lower costs. If the production costs become an issue Nvidia will focus on lowering that in the future although this is largely out Nvidia's hands because fabrication costs are just soaring. Unfortunately, the easiest ways to reduce production costs is to reduce VRAM and maintain production on older fab processes for longer, exactly what we have see with Turing. The other way is to simply be less innnovative and trickle out smaller changes each GPU, and do a load of rebrands with minor tweaks to clocks - exactly what AMD is doing.
Soldato
- Joined
- 14 Aug 2009
- Posts
- 2,755
No, it wouldn't. It's a new, better product, replacing the old one. It helps grow the PC gaming install base and AMD's share.
And let's not forget that they've jumped to $550 area with the 7970 when they were first on the market (never went back to $350-$400 area) and got the performance lead. At least they were giving their best, not like nVIDIA selling previously mid end cards from gtx 680 onward as high end cards and charging accordingly. Of course, people were happy to pay for that! In a way, nVIDIA's move to not compete with their best at "reasonable" prices (just as Intel stagnated with performance all these years), was perhaps lucky for AMD as it would have (most likely) completely remove them from the market - bad for us as well. Imagine de gtx680 selling for $249 and beating AMD's best which went for $549.
If Vega 64 an Vega VII could compete with 1080ti and 2080ti, they would have increased the price accordingly and Vega VII would have been $1000 at least. Of course, some would have defended this as HBM2 costs, 16GB vRAM costs, etc.
I'd say is reasonable to expect that. All games are targeted in terms of visual quality and gameplay at a certain HW which is, of course, what consoles and PC offer in that performance bracket. How much of a visual improvement is between the top settings and medium-high? Not much, if any during normal gameplay. If I turn everything up in GTA 5 1080p, on a 2080 it drops under 20fps if I remember correctly, but if you fine tune the settings, you can play up 4k@60fps and perhaps beyond, with similar IQ. Metro Exodus in 1080p, with extreme settings and all else turned up (NO RT), and still drops in certain areas under 60fps if I remember correctly. So basically, fine tune the settings, optimize for what gives back the best image quality for performance used, plus some form of upscalling like you said and 4K@60 isn't impossible in that scenario.
Fair play to that, I stand corrected on the performance scaling it was more than I remember. Back in those days Ati were superior at engineering onto a new process and used a far smaller die, Still though with Fermi's 6 month delay and the huge die size, I still think the 5850/5870 was priced fairly even at the extra £100 or so.
There was the 2gb version with extra display eyefinity6 gimmick and hdmi sound video decoding features etc.
Associate
- Joined
- 21 Apr 2007
- Posts
- 2,485
I'm not sure about that though current predictions are what Sony using Navi in some form for the PS5 and Navi 20 next year delivering Radeon VII + 10% @ $500. Even if we just write off 2019 what we are actually seeing is 2020 being a damp squib too. Even for consoles this isn't great because their hardware is teeing up to 4k its not exactly smashing through it for yrs to come or providing a lasting solution for VR/HMD, they are basically inserting a fill in product on that journey.
I dunno I just feel consumers are being milked to the extreme in this case I can see why AMD is doing what its doing and likewise for Nvidia and the whole thing stinks.
Soldato
Nvidia are milking, AMD are recovering from bankruptcy. It's not like AMD can smash out generation-defining architectures every year when their R&D budget is fractional compared to Nvidia as well as working back $1B in debt. Not sure how Sony are milking things either; consoles are designed to have a lifespan of many years, and what's going into the PS5 is light years ahead of the PS4.
Performance isn't going to scale quite like that between PC & Consoles. The latter will always be fine tuned to get the most out of hardware. All those games that see huge gains favouring AMD but are rare? That's what happens on consoles with everything, at least from first party studios. The hardware in the next consoles won't really be above a Vega 64 in performance, but that's more than enough for 4K 60. Of course, that means the consoles won't be pushing extreme settings which have disproportionate performance impact compared to visual fidelity (e.g. I always say this but the volumetric clouds on ultra from Odyssey vs very high; but we could just as easily point to other specific tech such as HBAO+ or HFT Shadows, or simply to other implementations of various settings which end up tanking performance for little visual difference when you move that slider an extra step). Most people don't bother tinkering with settings, nor do most reviewers test, but there's a LOT of performance gains for little visual difference to be had, even on PC. See for example Hardware Unboxed's game optimisation videos (eg: 20-30% performance improvement for no real visual decline in SotTR.
They're not getting VII + 10%. They're getting a V64(ish) equivalent system BUT that's being properly utilised, which is perhaps even more important. Of course, you don't get to tweak anything and so on, so in the end it's not really like getting a V64, but hey - pros & cons.
Last edited:
The fact it was close to the 295 is impressive.
Will multi GPU cards make their way back do you think?
Soldato
- Joined
- 28 May 2007
- Posts
- 10,061
I can't see sli/Crossfire as we know it making a big return. The future looks to be in small chips working as one which is multi gpu, Basically a little like what threadripper and zen 2 are. It's harder to do on gpu's but that's definitely the way it's going.
Soldato
Shouldn't many small chips working as one make multi-card easier to do, though?
I think mGPU will make a comeback, once DX12 gets more entrenched as an ecosystem. But that will only happen with the next console switch, so it's going to be a few years from now until we see it happen.
I think people overstate how bad CF/SLI is, plenty of great games work beautifully with it. I don't see why someone who would want to pay double for 20-30% more performance (1080 ti -> 2080 ti; V64 -> VII) is making such a better decision than someone who would rather get another card and enjoy >50% performance improvements in 1/2 - 2/3rds of relevant games.
I think people overstate how bad CF/SLI is, plenty of great games work beautifully with it. I don't see why someone who would want to pay double for 20-30% more performance (1080 ti -> 2080 ti; V64 -> VII) is making such a better decision than someone who would rather get another card and enjoy >50% performance improvements in 1/2 - 2/3rds of relevant games.