Charlie's thoughts on the fermi derivative parts

Page 3 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Soleron

Senior member
May 10, 2009
337
0
71
I think you are still missing the point. nV doesnt NEED to win price/performance. In Q4 2009, with nothing to compete against 5XXX in the "high-end", they still turned a nice profit.

You are comparing ATi and nV as if they are on equal footing. nV has better marketing, better brand recognition, and people are willing to pay for it.

480 and 470 + whatever cut-down chip comes out of Fermi will extremely well.

Charlie said that was from the 40nm GT200 cards (210, 220, 240) which had little AMD competition in that financial quarter. AMD released competitive parts (Redwood and Cedar) days after that earnings call. And they've made it nonsensical to buy the GT200 parts.

So that profit won't carry through to this quarter.

Charlie did say that that quarter was going to be good for NV well before the figures came out, so it didn't upset his prediction record or anything.
 

OCGuy

Lifer
Jul 12, 2000
27,224
37
91
For now. 3DFX was more established and had loyal fans willing to pay more for less. Look where that got them. This is generation 2 where ATI has the better price/performance product, and even the dimmest bulbs are starting to catch on.

I don't think any of the NV top brass or stockholders are ecstatic with nv's position in the market or currently retailing products. Which is why I think those products will rapidly improve. If history is any guide then Fermi II should be absolutely awesome.

True. It can only go so long before the whole "mindshare" thing starts to shift.

They know they cant continue to just grow the Xtors and chip size, especially since TSMC is having troubles with the next node again.
 

OCGuy

Lifer
Jul 12, 2000
27,224
37
91
charlie said that was from the 40nm gt200 cards (210, 220, 240) which had little amd competition in that financial quarter. Amd released competitive parts (redwood and cedar) days after that earnings call. And they've made it nonsensical to buy the gt200 parts.

So that profit won't carry through to this quarter.

Charlie did say that that quarter was going to be good for nv well before the figures came out, so it didn't upset his prediction record or anything.

lol.

Did I really just read that?
 

v8envy

Platinum Member
Sep 7, 2002
2,720
0
0
Again, as far as resources go, there's not much keeping AMD from pricing a 5850 par with the highest end GF104 - they're price gouging right now due to lack of competition.

From a BoM standpoint, no, nothing is keeping AMD from lowering the boom on NV. From a production capacity standpoint, yes, TSMC is keeping AMD from flooding the market and absolutely shutting NV out.

If ATI lowers prices they'll simply be gouged up by etailers. With unlimited capacity you bet ATI could clean up by selling 5850s for $150 and 5870s for $200. At that point the questions of "8800GTS/9800GTX/GTS250 or 5850?" and "GTX480 or 5850 in trifire for $50 cheaper" could both be answered with "Do you have Downs or not?"

So yeah, I don't expect any more price movement in the next 3-4 months than we've seen in the last 6. The hole will remain.
 
Last edited:

Ben90

Platinum Member
Jun 14, 2009
2,866
3
0
I do have a question about the article. In it he mentions:
GF100 is horribly weak in pixel ops. It has only 64 ROPs (Render Output Units), and that lack of pixel power shows up at higher resolutions.

Doesn't Cypress only have 32 ROPs? I thought all ROPs were the same, unless I'm wrong.
 

badb0y

Diamond Member
Feb 22, 2010
4,015
30
91
I doubt many people care about how much money nVidia or ATi makes, I just care about which product will give me the most performance per dollar. I, as well as many people, will jump on whichever ship makes more sense at the time. During the G80 era it was nVidia, right now it's ATi, who knows 6 months down the line nVidia may be the top dog again.
 

cbn

Lifer
Mar 27, 2009
12,968
221
106
I think you are still missing the point. nV doesnt NEED to win price/performance. In Q4 2009, with nothing to compete against 5XXX in the "high-end", they still turned a nice profit.

You are comparing ATi and nV as if they are on equal footing. nV has better marketing, better brand recognition, and people are willing to pay for it.

480 and 470 + whatever cut-down chip comes out of Fermi will extremely well.

Yes, I realize Nvidia doesn't need to win on price and performance when 2/3 of its profits (but 1/3 of its volume) come from HPC?

Therefore Nvidia will only rarely want to lower prices on Gaming video cards (especially if yields on these big double precision units are terrible). When this happens ATI will keep prices elevated.

This is why I am wondering if Intel won't see this as an opportunity to enter the market. When high profit margins exist....won't this cause competitors to increase?
 

Genx87

Lifer
Apr 8, 2002
41,091
513
126
Yes, I realize Nvidia doesn't need to win on price and performance when 2/3 of its profits (but 1/3 of its volume) come from HPC?

Therefore Nvidia will only rarely want to lower prices on Gaming video cards (especially if yields on these big double precision units are terrible). When this happens ATI will keep prices elevated.

This is why I am wondering if Intel won't see this as an opportunity to enter the market. When high profit margins exist....won't this cause competitors to increase?

Sounds like Intel has a lot of problems. From a product that doesnt deliver to a lawsuit that may crush them.
 

Soleron

Senior member
May 10, 2009
337
0
71
Yes, I realize Nvidia doesn't need to win on price and performance when 2/3 of its profits (but 1/3 of its volume) come from HPC?

They need consumer cards to pay off the cost of developing the GPU though. Could they really survive as an HPC-only company?

Doesn't HPC care a lot about power dissipation too? If Nvidia said 225W TDP for Fermi Teslas and it's now 275W then they're unlikely to carry on with it.
 

BenSkywalker

Diamond Member
Oct 9, 1999
9,140
67
91
3DFX was more established

nVidia released more 8800 parts then 3Dfx released total parts. nVidia also sold more 8800 series parts then 3Dfx sold parts total. Memories have an odd way of doing that to people, but ATi is far more established in the minds of the gaming market then 3Dfx ever was and they trail nVidia by a lot.

This is generation 2 where ATI has the better price/performance product, and even the dimmest bulbs are starting to catch on.

I think anyone who runs benchmarks all day caught on a while ago. People who play games, not as much. When ATi refuses to test obscure off the wall little niche games like World of Warcraft with each driver release, let's just say their reputation hasn't been improving in the mass market gamer segment. The average buyer, even of an add in board, doesn't ever glance at a benchmark chart. Not so much as a peak. They also don't care about 38.6 vs 43.5 fps, but they do care about their game just working without issue. ATi needs to be a little bit more proactive in getting drivers out that work with games before they ship and making sure they don't release drivers that screw up things in games like WoW. Yes, nV has their own share of issues on the driver front but day one pretty much everyone, even the red team loyalists, expect nV to have better peformance on new games. Then you have those that play games like WoW, go in chat on WoW and ask what video card to buy, an ATi 5850 or GeForce 9600GT for the same price. You would likely be shocked at the responses you get. ATi may be winning mind share in the technology enthusiast segment, they aren't doing anything close to what nV is doing for the gamers market.

Another issue is bundles. In terms of overall value, nV tends to smoke ATi on that front. Go on the Egg and half the nV boards come with a free game promotional offer of some sort(no, I didn't even check, just the rolling average). I was looking in for a fill in part as at the moment I'm not playing anything remotely demanding but the wife needed a new board for a new build. I was looking at a 4850 and a 250 GTS. The 4850 at the time was actually slightly cheaper, but the 250 GTS came with a game I was going to buy for $50 anyway having the end cost be that the 250GTS plus game was going to cost me ~$30 less. On a $110 part, that is a rather major price advantage. Anyone can try and argue that those sort of bundles shouldn't be factored in to the price, but the fact that nV works out deals to get games included regularly with their parts also increases the perception that they are the gamers choice. It is certainly a function of nV's marketing, but if I have to pay their marketing guys an extra $20 so I end up saving $50 on a game, well that's something that doesn't take a lot of careful consideration.

I don't think any of the NV top brass or stockholders are ecstatic with nv's position in the market or currently retailing products.

As far as their stockholders are concerned, the last quarter saw nV post higher gross profits then ATi had total revenue. I think you'll find that their stockholders are quite happy with their peformance at the moment. For a lot of different reasons, nV is still in a vastly superior financial position then ATi and for that matter AMD overall over the last couple of years(although the settlement from Intel certainly helped AMD out hugely and tips that scale heavily in end effect).

If history is any guide then Fermi II should be absolutely awesome.

I would agree, but not due to any idea of shortcomings in the design of Fermi from a business perspective. Mainly that conclusion comes from looking at exactly what Fermi is, where its' limitations are and how it will end after a die shrink. In terms of financial impact on nV, I've been hearing wait until next quarter, it is going to be terrible for the last nine months. So far, they have improved markedly over that time. Arguing the effectiveness of the engineering behind Fermi is something that can easily be done, so far those with red team leanings are hitting 0% on financial predictions for nV since the launch of the 5xxx series parts.

This is why I am wondering if Intel won't see this as an opportunity to enter the market. When high profit margins exist....won't this cause competitors to increase?

Intel can't compete. Based on the latest versions of Larrabee that were still in prototype stages they wouldn't be able to go toe to toe with a 9600GT/4770 without getting embarassed. Sure, there are a lot of price points up for grabs right now. Unfortunately filling them with parts that aren't much better then integrated isn't going to do much for you. When they first enter the market they need to win the tech enthusiast and OEM markets first before they have a chance at the gamers market. Running years old versions of the Doom3 engine at less then 10fps isn't going to get you there :)
 

Madcatatlas

Golden Member
Feb 22, 2010
1,155
0
0
As far as most are concerned Intel is one gargantuan bag of money. Money can get you anything (yes even love...)
I wouldnt be surprised if we see a midlevel price/performance effective card from them in the near future (5 years?)
 

Nemesis 1

Lifer
Dec 30, 2006
11,366
2
0
Oh so 480 is 500 bones ?

That is actually good considering I got my ubbber ATI X800 XT PE for 600 bones.

I do however think CPU and GPU prices are whack. IMO , a CPU should cost more then your GPU,,,, take it from there and see what you can get, Don't buy budget cards 57xx blah blah.. their budget cards. Get a 5870 or 5970 I believe is the model..

nVidia is drunk, their trying to compete with a dual GPU ATI card with their single GPU Furbi. Wait til the GTX 580 comes out,, dual gpu and will be 500 bones probably. Waiting is the key to success lol

Please Can you show all of us here were the x800xtpe sold for $600 . I bought 20 of them and that price your given is off the wall. $600 . I can't imagine anyone dumb enough to pay that price. Any sort of link . Nv last qt report and the one befor that was posted here. Do dig up the first one . See what I had to say . Clearly I stated that NV for those reports to be correct. Had to stop getting fab parts. Were waiting till the next report befor taking action with Sec . But NV never said in those reports they stopped getting fab parts . Which should have been in the report and clearly stated as such . Other wise they were decieving the share holders And I do own 10 shares LOL. Were ready to take action but we need to see next report 1st.

Wait until the SI or NI comes out it should spank the 580 and it should be out befor the 580.

I can't believe some fool said the high end doesn't matter.

What happened with the NV/ apple thing anyway??
 

Nemesis 1

Lifer
Dec 30, 2006
11,366
2
0
The 4850 was 260mm, on a 256bit bus and sold for ~$100. I think nV can build a chip with comparable build costs and sell it for twice as much and do OK with it. Another thing to keep in mind most people seem to be forgetting, the 5830, 5770 and 5750 all suck. They are grossly inferior to the parts they replace at each given price point- it isn't like nVidia has to come close to hitting one out of the park to be competitive.

One Charlie comment I have to point out-



The 480 is selling for $500 and he's saying they can't make a profit? That is so utterly moronic it has to be called out. If a magic fairy was delivering 5830 chips to ATi they would be taking a loss on them using that same logic. A $250 price rift between GPU cost and retail is huge.



The release of the 5xxx parts seemed to boost nV a lot on the financial end.

http://finance.yahoo.com/q/is?s=nvda

Having a price premium during a global recession may not be ideal, but they rebounded from it pretty quickly.

Please stop your killing me with laughter. If you could please tell all of us . What kind of yields NV is getting . For your statement to be true. You have to know NV yields on those parts . So please tell us . I agree with Charles if and only if yields are low . But if you know something we don't please enlighten us with a link telling us yields are just fine . Other wise your suger coating.
 

v8envy

Platinum Member
Sep 7, 2002
2,720
0
0
nVidia released more 8800 parts then 3Dfx released total parts. nVidia also sold more 8800 series parts then 3Dfx sold parts total. Memories have an odd way of doing that to people, but ATi is far more established in the minds of the gaming market then 3Dfx ever was and they trail nVidia by a lot.

My first add in board was an ATI all-in-wonder. 512K seemed so huge at the time. But ATI has only had credible 3d gaming hardware starting with the Radeon series. The Rage series was always an "also ran", it wasn't on the same level as the Voodoo products in the minds of enthusiasts.

Yes, the add in video board market has grown since 2000. And NV is in many more markets than 3DFX ever was. But at the time, they had a bigger % of the add in 3d board market than NV does today.

The average buyer, even of an add in board, doesn't ever glance at a benchmark chart. Not so much as a peak.

You make good points, but what else can we possibly use to project likely purchasing behavior? If we assume an irrational, uninformed and whimsical customer we would conclude NV could just as easily sell the GT210 for $600 by relabeling it as a GTX680. Since that isn't happening I'd have to conclude that at least some customers are capable of doing value analysis.

I think you'll find that their stockholders are quite happy with their peformance at the moment. For a lot of different reasons, nV is still in a vastly superior financial position then ATi and for that matter AMD overall over the last couple of years(although the settlement from Intel certainly helped AMD out hugely and tips that scale heavily in end effect).

Don't know about all the stockholders, but analysts aren't quite as bullish. http://finance.yahoo.com/q/ao?s=NVDA shows many more downgrades than upgrades. One to a sell, even. Ouch.

The chart shows a clear up trend over the year -- but then so does QQQQ.

When they first enter the market they need to win the tech enthusiast and OEM markets first before they have a chance at the gamers market. Running years old versions of the Doom3 engine at less then 10fps isn't going to get you there :)

I'm half joking when I say this, but didn't we just go over how NV is ruling the roost with inferior hardware through sheer force of marketing alone? If Intel was able to foist P4 and GMA off on the general public what's to stop them from marketing LarryB into a raging success story?
 

OCGuy

Lifer
Jul 12, 2000
27,224
37
91
I'm half joking when I say this, but didn't we just go over how NV is ruling the roost with inferior hardware through sheer force of marketing alone? If Intel was able to foist P4 and GMA off on the general public what's to stop them from marketing LarryB into a raging success story?

Well it depends on your definition of inferior. GT200 was not inferior to 4XXX performance wise, nor is Fermi going to be inferior to 5XXX, except on a dual-gpu hard-to-find niche board that goes for $650+.

Price/performance? Yes it will, but so was GT200, and it sold just fine.

Power usage/heat? Yes it will, but how many people actually care? Yes many people have come here to say they care, but they also said that about the GT200 when it came out, and it sold fine.

Your P4 example even takes it a step further....inferior performance AND inferior power/heat.....
 

cbn

Lifer
Mar 27, 2009
12,968
221
106
I'm half joking when I say this, but didn't we just go over how NV is ruling the roost with inferior hardware through sheer force of marketing alone? If Intel was able to foist P4 and GMA off on the general public what's to stop them from marketing LarryB into a raging success story?

Speaking of "marketing", maybe there is a better way to represent video card performance than FPS alone?

The latency issue you mentioned got me thinking about AFR vs SFR. Apparently SFR results in lower FPS, but lower latency.

Does this mean we might see ATI push SFR and synchronized GPUs as a way of fighting Nvidia's single GPU performance lead? Could this let ATI claim both higher FPS and lower latency at the same time using smaller multiple GPUs? Could such a strategy blend well with "Fusion" APU?
 
Last edited:

mooncancook

Platinum Member
May 28, 2003
2,874
50
91
Is ATi getting paid retail price for the 5830? If anyone thinks a company isn't turning a profit selling a video card for $500 using a chip that costs $250 they do not understand this market at all.

Well, depends on what the cost of this $250 includes. I know in a restaurant if the raw material cost of a dish is $2, and you sell the dish at $6, you can still end up with a net loss if you can't make/sell enough of the dish because running cost is high. If the $250 cost already include the necessary R&D and marketing cost then my bet is that there is profit.
 

v8envy

Platinum Member
Sep 7, 2002
2,720
0
0
And yet, I can still buy ATI 4 series and G92 hardware in bulk while all but trace supplies of GT200s are long since a distant memory. I don't think the GT200 was as much of a success as you believe.

I made the mistake of skipping the last video card generation and now I'm stuck skipping yet another one.
 

BenSkywalker

Diamond Member
Oct 9, 1999
9,140
67
91
Yes, the add in video board market has grown since 2000. And NV is in many more markets than 3DFX ever was. But at the time, they had a bigger % of the add in 3d board market than NV does today.

Actually they didn't, ATi dwarfed 3Dfx even in the add in board segment. 3Dfx purchased STB to better compete with ATi(they were significantly larger then nV too at the time). The tech enthusiast mindset was entirely about 3Dfx bad in the late 90s, but the larger market bought ATi by the truckloads.

Since that isn't happening I'd have to conclude that at least some customers are capable of doing value analysis.

The FX5200 sold a ton ;) I'm not saying that having an inferior product is the best way to conduct your business, but the only issues with nV's GPUs that we are discussing here are heat and noise, not performance and considering this is nV we are talking about, not even price.

Since that isn't happening I'd have to conclude that at least some customers are capable of doing value analysis.

Most customers, for anything, shop at a price point, not a performance metric. AMD dominated Intel for quite a few years in terms of absolute performance, performance/dollar and pretty much any other metric you can use, and never hit 33% market share. Sure, they had close to 100% of the enthusiast market, but in end effect they were still stuck in budget crap PCs from most companies and that perception remains today. The perception that GeForce is the brand to get in gaming circles is quite strong, hit the gaming sites and check out the forums and you'll find the less technical the site, the less relevant ATi becomes. I'm not saying that is the way things should be at all, just that is what we are looking at right now. Better marketting and focusing in on the types of things people bash nV for on these forums(TWIMTBP is *huge* for nV) would go a long way into helping ATi gain more mass market penetration.

Don't know about all the stockholders, but analysts aren't quite as bullish.

Look closely at that chart and in the last six months the pros outweigh the cons, only one downgrade. Back when nV was going toe to toe in every segment is when they were getting the bad ratings. In a huge economic dowturn, luxury items fall off much sharper then the broader markets, nVidia makes luxury items(their are cheaper alternatives in every segment).

I'm half joking when I say this, but didn't we just go over how NV is ruling the roost with inferior hardware through sheer force of marketing alone? If Intel was able to foist P4 and GMA off on the general public what's to stop them from marketing LarryB into a raging success story?

Hehe, I know you are kidding but can you imagine the RMA's they'd be dealing with when everyone got their new parts installed and they couldn't run anything made in the last three years over 10fps :p Most people don't care much about framerate, but it's like a car and most people not caring about 0-60 times, while they don't really care, they would at least like 0-60 to fall under the 'capable' category ;)

Well, depends on what the cost of this $250 includes. I know in a restaurant if the raw material cost of a dish is $2, and you sell the dish at $6, you can still end up with a net loss if you can't make/sell enough of the dish because running cost is high. If the $250 cost already include the necessary R&D and marketing cost then my bet is that there is profit.

You are very correct in how the part could end up being a major loss for them. They need to sell a lot of parts to recoup R&D- this R&D is spread out across an entire product line, so it isn't like they need to sell $1Billion worth of just the GF100 parts, but they do need to sell an awful lot of GF1xx based parts to recoup their money. The issue I have with what Charlie is stating is that the part as it is made is unprofitable. The only way it becomes unprofitable is if it doesn't sell enough units. His claims that there will only be on batch of these cards makes no business sense. Much like the $2 dish, you don't start turning people away after the first few meals are served or you will lost a ton based on your overhead. This is the only way nV loses money selling a $250 part for $500, if they don't sell a lot of them. Gross margins need to remain higher then R&D expenses to keep them profitable, it's as simple as that. That raw margins on the boards, while certainly having an impact, isn't nearly the factor that scales of economy are due to the nature of the industry and how much of your total product cost is in R&D.

And yet, I can still buy ATI 4 series and G92 hardware in bulk while all but trace supplies of GT200s are long since a distant memory. I don't think the GT200 was as much of a success as you believe.

Really look at it and the GT200 was a "G92.5" part. Feature set was almost identical, the GT3xx parts actually have a larger feature rift then the GT2xx and G82 parts.
 
Last edited:

Ben90

Platinum Member
Jun 14, 2009
2,866
3
0
Speaking of "marketing", maybe there is a better way to represent video card performance than FPS alone?

The latency issue you mentioned got me thinking about AFR vs SFR. Apparently SFR results in lower FPS, but lower latency.

Does this mean we might see ATI push SFR and synchronized GPUs as a way of fighting Nvidia's single GPU performance lead? Could this let ATI claim both higher FPS and lower latency at the same time using smaller multiple GPUs? Could such a strategy blend well with "Fusion" APU?

Until we see a nice single frame rendering solution, I would rather have the hot expensive GTX 480 as I try to keep my input delay to a minimum. If we could see a cheaper implementation of Hydra, or even better, a native solution from ATI/NVidia then I would be very inclined to start looking at the more cost effective Crossfire/SLI solution using mid-highrange cards.

Once a SFR solution is mastered I believe ATI's current architecture will whoop NVidia's current pretty hard.

I don't understand why ATI/Nvidia cant do what Hydra does 100% in software. Yea it would take the driver team a while to implement, but the payoffs would be great.
 
Last edited:

v8envy

Platinum Member
Sep 7, 2002
2,720
0
0
I don't understand why ATI/Nvidia cant do what Hydra does 100% in software. Yea it would take the driver team a while to implement, but the payoffs would be great.

Brilliant! Let's spend lots of development resources so the enthusiast can buy two $120 cards or four $50 cards instead of a $500 card. We'll make a fortune!

Both NV and ATI are walking a narrow line. It's in their best interest to make you desirous of multiple high end cards, yet at the same time make it non-viable to build a redundant array of entry level cards. AFR fits the bill. SFR not so much.
 

cbn

Lifer
Mar 27, 2009
12,968
221
106
Both NV and ATI are walking a narrow line. It's in their best interest to make you desirous of multiple high end cards, yet at the same time make it non-viable to build a redundant array of entry level cards. AFR fits the bill. SFR not so much.

Yes, but ATI is not making the Big single GPU video cards and on top of that they have compatibility with Fusion to consider.
 
Last edited: