Prediction - Intel to buy Nvidia?

['Wes]

Drunkenmaster likes to construct a post that makes him feel better about his purchase.

 

[oweneades]

Really now? Massively slower than what exactly?

You have to remember that these are the words of a self confessed AMD share holder who claims to have over 20K worth of AMD stock. I'd argue that it is hard to make a subjective post (not that they even come close) in such circumstance and are not the words of somebody I would trust for impartial analysis of the industry. Each post always paints ATI in the best possible light, and Nvidia in the worst. It gets tiresome after a while.

IIRC Drunkenmaster has posted before that the GF104 is both larger and more expensive to produce than the Cypress GPU in the HD5850/HD5870.

This is most likely what he is reffering to in the post above.

 

[Tombo]

Drunkenmaster likes to construct a post that makes him feel better about his purchase.

He makes big posts about everything, not really anything to do with his purchases.

 

[Freddie1980]

Drunkenmaster likes to construct a post that makes him feel better about his purchase.

As do 90% of people around here as spending big money on hardware makes some of us insecure.

 

[An Exception]

I think drunkenmaster is an ATI/AMD fanboy, all I get from reading his posts is how nvidia are terrible and how they are going under.

...right.

No he's not an AMD fanboy, he just hasn't got much love for Nvidia which is why he bothers to post about the issue which is Nv.

This however doesn't mean his points should be discounted with grounds of Bias, they are very logical and plausible analysis's similar to Charlie.

"...right" just sounds like you have blind faith all will be OK because it just will. Perhaps while trying to discredit him you could actually try finding fault with his argument if you feel up to it.

Lot's of GPU companies hit the wall before Nv, so it's not like it's beyond the realms of possibility, Do you remember 3DFX?

Nv are now at cross roads in their story, and they have between now and until their cash reserves run dry to get their **** together, what they do in the next 12 months will make or break them.

 

[An Exception]

Lets pick one of the easier points to shoot down in flames.



Really now? Massively slower than what exactly?

Hmmm, from that comment I would guestimate you need to do a little more research before bringing counter arguments to the table, else you won't be shooting any points down in flames.

GF104 has larger die than a 5870 but sell for how much compared to a 5870?

 

[mattyfez]

To say Intel and NVIDIA aint exactly best friends is an under statement! NVIDIA would have to be on their knees before they sell up.

NVIDIA wants in on the CPU market so it can produce full hardware platforms, and has been pretty much blocked out by the CPU giants. Hence their massive drive toward cuda and other gpgpu type stuff.

Also NVIDIA is leagues ahead of ATI in the pro workstation arena. It depends if their expensive spurt in parallel processing takes off in the long run. It could go either way, but its worth noting that its something Intel tried, and failed miserably, to do with Larabee or whatever it was called.

 

[Marvin]

To say Intel and NVIDIA aint exactly best friends is an under statement! NVIDIA would have to be on their knees before they sell up.

Don't forget Intel could simply buy all of Nvidia's shares in a hostile take over. It doesn't have to be agreed.

 

[layte]

Hmmm, from that comment I would guestimate you need to do a little more research before bringing counter arguments to the table, else you won't be shooting any points down in flames.

GF104 has larger die than a 5870 but sell for how much compared to a 5870?

the gf104 is massively slower than cores that are smaller and sold at a higher cost

So which competing part is it massively slower than?

If I wanted to read AMD propaganda I would visit Semiaccurate.com.

 

[layte]

double post..

 

[An Exception]

So which competing part is it massively slower than?

If I wanted to read AMD propaganda I would visit Semiaccurate.com.

http://www.anandtech.com/bench/Product/162?vs=180

The performance per mm2 of GF104 is pretty ugly considering it's actually larger than the 5870, however Nvidia have given it a very attractive price, hence why they are probably not making money on it.

 

[Rroff]

See you said this, for a long time, every time Larabee was brought up, however their Sandybridge GPU shows a MASSIVE, monumental leap forwards in low end gpu power, also most importantly, Anand who used to mock Intel IGP's for giving awful IQ and dodgey drivers suggest their new GPU is massively massively better and giving equal IQ aswell as performance to a 5450.

Lets be honest, theres VERY little difference between a small 12core Sandybridge gpu and sticking one with 10 units stuck together on a discrete card.


They've clearly addressed, significantly, IQ, power, ability to deliver it and considering the supposed power of Sandybridge chips, their 12core gpu seems set to use very little power, to the degree that you could stick 20 of these units together and have a workable product.

I mean a Fermi high end and low end, a Nvidia IGP or AMD IGP is little more than just one or two units of a high end card, the architecture is all but the same.

As I've said - developing great hardware means nothing on its own - nVidia's latest generations have been fairly ugly hardware wise but they've still been somewhat sucessful. Intel have a far too cut and dried approach to their products which works really well for things like CPUs and peripherals, etc. but doesn't work at all for performance GPUs.

Also no, no one said anything similar at the FX launch, no one predicted Nvidia's downfall or their unwillingness to change their ways, it was one bad product, with a good one before and after it. the 280gtx through to the current gen ALL show the same signs of sticking to certain design philosophy that has made them progressively worse over 3 years.

Granted the scenarios aren't identical, but there was a lot of doom and gloom over the FX launch with people predicting how nVidia was on its last legs and about to go the way of 3dx...

 

[layte]

http://www.anandtech.com/bench/Product/162?vs=180

The performance per mm2 of GF104 is pretty ugly considering it's actually larger than the 5870, however Nvidia have given it a very attractive price, hence why they are probably not making money on it.

So which competing part is it massively slower than?

I'm picking up on the massively here as it's not, even when compared to the 5870. Well IMO of course, but as the saying goes, opinions are like a..... everyone has one.

I'll not argue that when it comes to things like performance/mm2 Nvidia's solutions leave a lot to be desired, but only a fool would think that this will be an unchanging situation. They still have a lot of very smart people working for them, and if AMD can bring things back from the brink of debtageddon (with the help of $1.2Bn in Intel hush money) then why do some persist in constant doommongering?

The very gf104 we are questioning gives an indication of what Nvidia have planned. Given that it is a much smaller core than the gf100, just how does it's performance per transistor stack up.....?

 

[Rroff]

They still have a lot of very smart people working for them, and if AMD can bring things back from the brink of debtageddon (with the help of $1.2Bn in Intel hush money) then why do some persist in constant doommongering?

To be fair AMD have some very very good business people, they wouldn't have survived otherwise. nVidia has some smart people but they don't tend to be at the same level of calling the shots and gimped by the higher end of the company (IIRC a good number have left to work for intel or IBM can't remember which within the last few months). I don't think nVidia is going down any time soon tho - they've worked hard to tie themselves into the video gaming industry and pushing CUDA as an industry standard both of which seem to be paying off slowly.

 

[Freddie1980]

Don't forget Intel could simply buy all of Nvidia's shares in a hostile take over. It doesn't have to be agreed.

There public listed but what we don't know is who owns the shares. I would have thought Dear Leader Jen-Hsun Huang would still own a sizeable portion of the company which he would have to agree to sell before any take over would be successful.

 

[PaulyD]

Indeed, even though public listed someone will have a commanding share

 

[An Exception]

I'll not argue that when it comes to things like performance/mm2 Nvidia's solutions leave a lot to be desired, but only a fool would think that this will be an unchanging situation. They still have a lot of very smart people working for them, and if AMD can bring things back from the brink of debtageddon (with the help of $1.2Bn in Intel hush money) then why do some persist in constant doommongering?

Because the perf/mm2 isn't the only issue Nv has, they have a whole storm of challenges about to arrive at it's front door, namely loosing most of it's current revenue stream (low end).

But also many other things for example like surviving the 40nm downpours.
If Amd's 6000 series is even more effiecient per mm2 then they can afford to squeeze GF104 into a loss making product while staying just as profitable as it is now. GF104 only needs to make a slight loss per unit to have catastrophic effects on Nv's cashflow far more damaging than the GTX465/470 etc. due to the fact it's a high volume part and the losses are much more scalable.

Nvidia would then be forced to reduce production of GF104 to create artificial demand in order to keep prices above Amd's and therefore profitable (or break-even), this unused TSMC allocation will likely then be handed to Amd.
The above means Nv is going to be unlikely to recover market share and lose allot of clout with regards to putting weight, cash & resources aside to influence developers to optimise for Nv architectures.
This will put further downward pressure on Nv pricing and exacerbate their losses or they will have to further scale back production which then turns into a viscous cycle.
It get's even worse for Nv if we take GF106 etc into account as they will have the same problems as GF104.

If Nv can produce a Gf104 based GTX475 in high enough Qty's due to increasing yields so there are more 475 being produced than 460's then this may be a band-aid for Nvidia's bottom line, that is currently lined up against a wall.

The biggest saving grace for Nv is the professional market, but there has been talk of Amd's intentions to push hard in this Nv stronghold, the VERY last thing Nv wants is a price war in this market like when RV770 launched in the gamer market.

There are many other chinks in the armour but to continue would just be too doom and gloom even for a realist.

 

[An Exception]

Indeed, even though public listed someone will have a commanding share

Yep that's right, but the public will own the most shares, so Intel would simply have to wait for the public to cash in their shares and keep snapping them up until they then have a commanding share. Of course Intel wouldn't be interested in doing this.

 

[layte]

Because the perf/mm2 isn't the only issue Nv has, they have a whole storm of challenges about to arrive at it's front door, namely loosing most of it's current revenue stream (low end).

But also many other things for example like surviving the 40nm downpours.
If Amd's 6000 series is even more effiecient per mm2 then they can afford to squeeze GF104 into a loss making product while staying just as profitable as it is now. GF104 only needs to make a slight loss to have catastrophic effects on Nv's cashflow far more damaging than the GTX465/470 etc. due to the fact it's a high volume part and the losses are much more scalable.

Nvidia would then be forced to reduce production of GF104 to create artificial demand in order to keep prices above Amd's and therefore profitable (or break-even), this unused TSMC allocation will likely then be handed to Amd.
The above means Nv is going to be unlikely to recover market share and lose allot of clout with regards to putting weight, cash & resources aside to influence developers to optimise for Nv architectures.
This will put further downward pressure on Nv pricing and exacerbate their losses or they will have to further scale back production which then turns into a viscous cycle.
It get's even worse for Nv if we take GF106 etc into account as they will have the same problems as GF104.

If Nv can produce a Gf104 based GTX475 in high enough Qty's due to increasing yields so there are more 475 being produced than 460's then this may be a band-aid for Nvidia's bottom line, that is currently lined up against a wall.

The biggest saving grace for Nv is the professional market, but there has been talk of Amd's intentions to push hard in this Nv stronghold, the VERY last thing Nv wants is a price war in this market like when RV770 launched in the gamer market.

There are many other chinks in the armour but to continue would just be too doom and gloom even for a realist.

You make some very good points. Lots of use of the word if to paint a possible worst case scenario but still some interesting points.

I'll raise the issue of the pro market. It will take much more than a price war for AMD to break into NV's market share. Most of what they currently release is nothing more than a warmed over gaming part with a much higher price tag. Where are all the certified drivers? Where are the engineers working with end users and software vendors? It takes more than just price to break into a market ruled by reputation, and at the moment that is something AMD have very little of.

 

[bridder]

You make some very good points. Lots of use of the word if to paint a possible worst case scenario but still some interesting points.

I'll raise the issue of the pro market. It will take much more than a price war for AMD to break into NV's market share. Most of what they currently release is nothing more than a warmed over gaming part with a much higher price tag. Where are all the certified drivers? Where are the engineers working with end users and software vendors? It takes more than just price to break into a market ruled by reputation, and at the moment that is something AMD have very little of.

As I understand it though the bulk of NVs profilt and AMDs from a GPU perspective comes from the bottom end of the graphics market, please correct me if I am wrong? As I genuinely don't know. I suppose what I am saying is from a business perspective the pro market and high end gaming isn't going to sustain a company the size of NV?

Lets face it they have been having a mare of it really, they EOL'd the 200 series 6 months before they had a replacement because they simpy couldn't compete with AMDs 5000 series, that is a truly awful indictment of just how badly flawed the 200 series was, from a business perspective, ignoring the fact that it once again (as always seems to be with NV) allowed NV to keep hold of the single fastest chip crown, there is no point in hanging onto the crown if it is going to bankrupt the company.

Having said that taking away their last quarter results they have continued to make a profit though (forgive me I'm thinking out loud here...), unless Charlie is to be believed and they were indeed channel stuffing, and that will have its own consequences again from a business point of view re the 400 series...

Half a billion on bumpgate (think about that simple word for a minute BILLION!) so far! Poor relations with virtually every console manufacturer on the planet because of their unswerving belief that they are superior, better etc and what sounds like a kick out at Apple and they have all the makings of a company heading only one way. Re the OP and the thread topic I don't think Intel want or need NV, I'm sure they have some absolutely fantastic engineering skills though but a buyout I cannot see. In fact yes I can see it, they will buy it up when the stock reaches junk prices.

I don't wish this, I just see it happening and I can tell you now I don't have 20K sunk into any company, maybe my house

random thoughts,
J.