"Too Little, Too Late....Too Expensive" Did you agree?

[OCGuy]

Tech boards were lit up with chatter during the Fermi launch, which was a few months behind its competitors offerings. It can be summed up by the linked article:



"We ran benchmarks in a variety of current titles and, on the whole, the Fermi cards narrowly outperformed their ATI equivalents....."

"....if we average all the results, Nvidia's edge looks to be between 5% and 10%."

"While being the proud owner of the "world's fastest GPU" is always a short-lived thrill, I feel that it'll be a shorter than usual thrill for Fermi owners. That 10% edge that the GPU has over ATI isn't going to last long."

"My advice ... keep hold of your money and wait and see."


http://www.zdnet.com/blog/hardware/...ittle-too-late-too-hot-and-too-expensive/7824



All signs are pointing to 290 being basically in the same situation, minus the power issues.

Did/do you agree with the headline in 2010? Would you apply it equally now?

Try to keep it as clean as possible, sighting facts, complete thoughts, etc. Full disclosure, I was one of the first regular consumers in the world with a GTX480, even though it was expensive, hot, etc. I will probably have a 290 as well.

 

[Stuka87]

All depends how fast the 290X is. nVidia and AMD both have to wait nearly a year for 20nm releases. There is a chance nVidia could release a Titan Ultra, that is basically a full blown GK110, as opposed to a cut down one.

But if 290X is faster, it most likely will be until mid next year.

And in the case of Fermi, it was what, just over 6 months behind AMD? For the 290X, it is a few months behind the 780, but not 6 months behind. I personally consider the Titan to be outside of the main release cards. Its an ultra expensive singular model. So i feel the 780 is more in line with what AMD is competing with.

 

[railven]

Well this thread isn't going to go far...

If the shoe fits, wear it. Regardless, you won't change peoples opinions on the team they are invested in, so not sure what you expect for results.

 

[Gikaseixas]

A little too early to start a thread like this given that nobody knows how much it will cost, how fast it will perform, how much it will consume and many other variables but if you insist let's start with price which seems to be lower:

R 290X (most likely) less expensive than current fastest card, GTX 480 was more expensive so your comparison is flawed already

 

[tviceman]

I think it is too late, and if it is MSRP $700 then definitely yes it's too little.

 

[Teizo]

Is this based on the fact that 20nm cards are around the corner?

I can say this...Based on AT's review of the R9 280X Toxic, the 290's are going to be a force to be reckoned with until they do.

 

[OCGuy]

All depends how fast the 290X is. nVidia and AMD both have to wait nearly a year for 20nm releases. There is a chance nVidia could release a Titan Ultra, that is basically a full blown GK110, as opposed to a cut down one.

But if 290X is faster, it most likely will be until mid next year.

And in the case of Fermi, it was what, just over 6 months behind AMD? For the 290X, it is a few months behind the 780, but not 6 months behind. I personally consider the Titan to be outside of the main release cards. Its an ultra expensive singular model. So i feel the 780 is more in line with what AMD is competing with.

I started a thread about who to blame for performance stagnation, and TSMC is at the top of my list. I guarantee the 290 being shrunk could have been packed with all sorts of extra goodies, but that isn't the reality we are working with.

Not counting Titan is up to you, but it is a retail launch card available to anyone with the dough or a credit card.....

 

[ArchAngel777]

I think it is too late, and if it is MSRP $700 then definitely yes it's too little.

If this thing is priced at $99, it is game over for nVidia.

 

[Kenmitch]

The R9 290 is currently overpriced and bring no bang for buck....Don't believe me then look at the gouging going on over at newegg. I didn't believe the extent newegg would go thru to rape early adopters! Wholly geepers batman!

http://www.newegg.com/Product/Produc...82E16814127758

Can't believe it's sold out currently!

Of course the entire above post is a joke

Without reviews it's not possible to judge the 290 series.

 

[Grooveriding]

nm

 

[ArchAngel777]

The R9 290 is currently overpriced and bring no bang for buck....Don't believe me then look at the gouging going on over at newegg. I didn't believe the extent newegg would go thru to rape early adopters! Wholly geepers batman!

http://www.newegg.com/Product/Produc...82E16814127758

Can't believe it's sold out currently!

Of course the entire above post is a joke

Without reviews it's not possible to judge the 290 series.

^^

Without reviews and factual information.... Well, most people already understand this.

 

[Teizo]

http://www.newegg.com/Product/Produc...82E16814127758

Can't believe it's sold out currently!

It even comes with a one year warranty!

 

[stahlhart]

Well this thread isn't going to go far...

Awww, why the pessimism? I'm sure that this will be end up turning out to be a nice, cordial and civil discussion.
-- stahlhart

 

[OCGuy]

Without reviews it's not possible to judge the 290 series.

I think we can take an educated guess, due to it being stuck on the same process. 28nm is much more mature now, but 5-10% over Titan/780 before overclocking would be impressive, it just would have been better for AMD had it happened in the Spring.

 

[Gikaseixas]

Without reviews it's not possible to judge the 290 series.

Thats my opinion too. How can one fairly compare if the NDA lifts only next week?

 

[Vesku]

Who can say in any meaningful way without knowing the performance and price?

 

[railven]

Because it's late, possibly marginally faster, and most importantly an AMD product trying to be put into the same negative light a prior NV product was in.

And this cycle will keep repeating as if someone is going to prove someone else's opinion wrong. /shrug

 

[OCGuy]

But if 290X is faster, it most likely will be until mid next year.

This is true....AMD and NV churned out refreshes very quickly that round. 290 will have a fairly long run at the top (again that is an assumption but it is likely) unless NV comes out of nowhere like they did with the 580. But there really isn't much to refresh on this node, so Maxwell will be the next entry to the market.

 

[Elfear]

I think we can take an educated guess, due to it being stuck on the same process. 28nm is much more mature now, but 5-10% over Titan/780 before overclocking would be impressive, it just would have been better for AMD had it happened in the Spring.

As you say, AMD is late to the game which takes some of the enthusiasm out of the launch (unless performance is simply mind-numbing). That said the big difference I see and others have pointed out is pricing. If the 290 comes with the same performance as a 780 and is a decent amount cheaper, it will be better situation than the 480 launch IMO.

 

[GaiaHunter]

Are we comparing the $1000 and $700 price bracket with the $500 market?

AMD had a full 40nm line up, including the 5970 before the GTX 480 arrived.

Also people were used to a higher pace of development then - but we were stuck with 40nm for a while and it seems 28nm will be here for longer.

 

[OCGuy]

As you say, AMD is late to the game which takes some of the enthusiasm out of the launch (unless performance is simply mind-numbing). That said the big difference I see and others have pointed out is pricing. If the 290 comes with the same performance as a 780 and is a decent amount cheaper, it will be better situation than the 480 launch IMO.

We actually know more about performance (based off leaked AMD slides and benchmarks that are starting to roll out) than we do about the market response re: pricing.

NV is not pricing tone-deaf.....it wasn't too long ago that my 2nd GTX280 cost me 40% less than the first, and NV had a Tier 1 revolt, causing a loss in XFX exclusivity. Who knows how much they had to pay EVGA to not start selling AMD........

 

[raghu78]

OP you are speculating about stuff which everybody is going to know in 5 days. anyway pricing is still unknown. so the too expensive part is yet to be determined. too little - if that is the case why are people still paying USD 250 -300 (60% - 70% higher) more for GTX 780 for a measly 20% better performance over GTX 770 when both are stock and 30 - 35% at max overclocks for both cards.

now about Fermi vs Evergreen vs Northern Islands , HD 5870 was a good chip when it released. But as an architecture Evergreen was not as good as Fermi for DX11 games. You could see that even in 2010 with games like Metro 2033. but by late 2011 when games like BF3 launched it was obvious that Fermi was a much superior DX11 architecture with tesselation and compute performance which embarassed AMD HD 5870 and HD 6970.

AMD never fixed the performance issues present in HD 6970 and HD 5870 cards with deferred rendering engines like Frostbite 2/3 . With 12.11 never settle drivers AMD fixed that for GCN and AMD HD 7000 cards.

http://www.pcgameshardware.de/Battl...-4-Beta-Test-Grafikkarten-Benchmarks-1090869/

GTX 580 destroys a HD 6970 and HD 5870 in games like BF4. this is what a forward thinking architecture is capable of. the GTX 580 is 3 years old now and an overclocked GTX 580 at 925 - 950 mhz can play max settings at 1080p in BF4 at 40 fps. :thumbsup:

The initial implementation of GTX 480 was more of an engineering issue. When Nvidia fixed the flaws and brought out GF110 aka GTX 580 Fermi kicked ass. AMD could not catch up even with Northern islands HD 6970.

Having said that with Southern Islands chip HD 7970 , AMD built a chip which performed well in DX11 games and with good compute performance. Nvidia though went the other direction and sacrificed compute for better gaming efficiency with GK104. with GK110 I believe Nvidia had to give up gaming efficiency for better compute performance.

So when Hawaii and GK110 are benchmarked in gaming and compute you will get a true idea of which is the superior architecture for both gaming and compute. I believe AMD has already won the perf / sq mm battle as almost everyone by now realizes that R9 290X is a Titan equivalent in perf but at a much smaller die - 438 sq mm for hawaii vs 551 sq mm for GK100. the other two aspects are perf/watt and perf/$ which we will know on Oct 15th. :thumbsup:

 

[Saylick]

"Too Little, Too Late....Too Expensive"

Too little? We don't know for sure until reviews are out, but the word on the street is that 290 ~780 and 290X ~Titan.

Too late? Considering the 780 was released around 4.5 months ago, I wouldn't call that too late. Too late would be releasing the card after holiday season, considering if you are only allowed to release a product one time during the year, you want to make sure it's released right before holiday season. It's a bit behind for sure but not far behind that the card is irrelevant. The big deal isn't the time between the 780 and Hawaii; the big deal is the time gap between Hawaii and next-gen nVidia cards. If nVidia doesn't release a superior product for another year - which is what's going to happen considering 20 nm-class cards won't be out until Q4 of 2014 at the earliest, that means AMD has 1 years time of market advantage. From a business perspective, 1 years worth of having the faster and possibly more bang-per-buck GPU is worth the 4.5 month delay imo.

Too expensive? We won't know the official price till launch, but rumors point to $499 290 and $650 290X, with the sweet spot being the 290 as it's spec'd to be a $500 GTX 780 equivalent with potentially better high res/AA performance due to higher bandwidth, OC potential, and ROP count.

Comparing all this to Fermi, you have:

Too little? Probably or probably not, depending on who you ask. Keep in mind that Fermi was broken from Day 1; GTX 580 should have been what Fermi was on launch, as the 580 commanded another ~15% performance boost over a 480 if I recall correctly.

Too late? Again, 6 months late is really only a big deal if you know that means it will only translate into a 9 month market advantage. More importantly, the 480 was released after Holiday Season of 2009, i.e. they missed a golden opportunity, whereas the 5xxx product line was released in Sept of 2009. The 6xxx product line was released in December of 2010, which means Fermi only enjoyed a 9 month long performance advantage at the cost of a 6 month delay. Compare this to a 4.5 month delay to a 1-year long market advantage.

Too expensive? I'd say so. Fermi ran HOT [redacted]. For what it was priced at, you could get a cheaper, less hot/power hungry 5870 which provided 85% of the performance of a 480. If Fermi wasn't broken on Day 1, i.e. if the 480 was in reality the 580, then I'd say the price premium could be justified, but that's not how things played out.

Warning issued for inappropriate language.
-- stahlhart

EDIT: My bad, didn't know that was pushing it too far for inappropriate language... :\

 

[railven]

Depending on power and performance metrics, this card might be more comparable to a GTX 680 than GTX 480:
Late (3.5 months vs 4.5 months)
Cheaper (R9 290X is most likely going to be cheaper than GTX 780/Titan)
Faster (possibly I guess, marginally faster ala 680 versus 7970.)
Power (if it uses less power than 780/Titan, well that's another notch more towards a GTX 680 comparative)

EDIT: And everyone seems to have loved the GTX 680.

 

[Gikaseixas]

Depending on power and performance metrics, this card might be more comparable to a GTX 680 than GTX 480:
Late (3.5 months vs 4.5 months)
Cheaper (R9 290X is most likely going to be cheaper than GTX 780/Titan)
Faster (possibly I guess, marginally faster ala 680 versus 7970.)
Power (if it uses less power than 780/Titan, well that's another notch more towards a GTX 680 comparative)

EDIT: And everyone seems to have loved the GTX 680.

You nailed it :sneaky:
This has indeed more similarities with the GTX 680 launch