Geforce4 Ti 4800!

[Trawler]

I haven't seen this posted here yet. My apologies if it has!

http://www.anandtech.com/showdoc.html?i=1741&p=4

Unfortunately timing for NV30 doesn't seem to be working out as perfectly as once hoped. Although there may be something to show off later this year, NVIDIA is telling their partners that chips will be available early next year.
...
For starters we have the GeForce 4 Ti 4800, but don't let the name excite you - the card is nothing more than a GeForce4 Ti 4600 with AGP 8X support. Even more disappointing is that there's talk of a GeForce4 Ti 4800-SE that runs at the same clock speed as the current-gen Ti 4400 but with AGP 8X support.

 

[Evildeus]

This part is intriging:

Luckily not all of NVIDIA's graphics announcements are along the lines of delays or the Ti 4800. NVIDIA was over today showing Matthew and I something very impressive that you'll be able to read about in the coming week(s).

What is it if it's not a NV30?

 

[martrox]

Well, Anand also has to beat on ATI once again:

For starters we have the GeForce 4 Ti 4800, but don't let the name excite you - the card is nothing more than a GeForce4 Ti 4600 with AGP 8X support. Even more disappointing is that there's talk of a GeForce4 Ti 4800-SE that runs at the same clock speed as the current-gen Ti 4400 but with AGP 8X support.

We can't only fault NVIDIA for misleading nomenclature though, the difference between ATI's Radeon 9500 and Radeon 9500 Pro is huge yet the difference in name implies nothing more than clock speed changes.

 

[JF_Aidan_Pryde]

Anand is getting desparate or what?

 

[nAo]

Well, Anand also has to beat on ATI once again:

It seems to me, from the text you quoted, he bashes Nvidia and ATI both.

 

[martrox]

Well, Anand also has to beat on ATI once again:

It seems to me, from the text you quoted, he bashes Nvidia and ATI both.

Key term: also

 

[nAo]

Key term: also

Yeah, unfurtunately your key term doesn't make sense with the nature of your post.But of course u'll deny that, I know. So we can just stop that here.

 

[Snap]

Yeah, unfurtunately your key term doesn't make sense with the nature of your post.But of course u'll deny that, I know. So we can just stop that here.

Given the context of Anands piece (products late - well behind promised dates, new products just more refreshes of two year old products) isn't it correct that NV are being bashed? How does this correlate to ATI when they have a complete line-up refresh with no discrete DX7 products there and most of them DX9 ready?

Doesn't your own whining demonstrate more about your bias than Martox's?

 

[martrox]

Key term: also

Yeah, unfurtunately your key term doesn't make sense with the nature of your post.But of course u'll deny that, I know. So we can just stop that here.

nAo,
Sorry, but I won't deny it kind of bothers me that Anand has to say something bad about ATI in an article about nVidia, and uses it to justify what nVidia is doing. Guess I would have has less of a problem if he has said something about the 9000 series, as it's naming is a bit more missleading(not DX9) than 9500 vs. 9500 PRO. I really don't feel that the difference in naming here "implies" that the only difference is speed. Is speed the only difference between the 9700 & the 9700 PRO? The point is that the reason Anand even brought up ATI was to justify nVidia, period.

 

[nAo]

Given the context of Anands piece (products late - well behind promised dates, new products just more refreshes of two year old products) isn't it correct that NV are being bashed?

Yes, it is. Have you read anything from me stating it's not?

How does this correlate to ATI when they have a complete line-up refresh with no discrete DX7 products there and most of them DX9 ready?

It just doesn't correlate. In fact no one is saying ATI has not the best lineup.

Doesn't your own whining demonstrate more about your bias than Martox's?

No. In fact my post wasn't meant to bash nvidia or ati, and it's pretty clear.

ciao,
Marco

 

[Joe DeFuria]

Actually:

We can't only fault NVIDIA for misleading nomenclature though, the difference between ATI's Radeon 9500 and Radeon 9500 Pro is huge yet the difference in name implies nothing more than clock speed changes.

I disagree with that statement, so I tend to agree with the assesment that Anand is trying to grasp at atraws here to somewhat justify nVidia's naming.

"9500 Pro" does NOT imply faster clock speeds to the consumer. All it implies is that it is faster (or at least a somehow better) product. I don't care if that's achieved through faster clocks, more pipes, more bandwidth, whatever.

On the other hand, a GeForce4 4800 SE?! One can argue that SE implies it is better than a NON-SE 4800. What does "SE" mean? (Special Addition?) Reminds me of the Intel 386 / 386-SX debacle.

 

[Evildeus]

This part is intriging:

Luckily not all of NVIDIA's graphics announcements are along the lines of delays or the Ti 4800. NVIDIA was over today showing Matthew and I something very impressive that you'll be able to read about in the coming week(s).

What is it if it's not a NV30?

"bump" Some information?

 

[Bjorn]

"9500 Pro" does NOT imply faster clock speeds to the consumer. All it implies is that it is faster (or at least a somehow better) product. I don't care if that's achieved through faster clocks, more pipes, more bandwidth, whatever.

On the other hand, a GeForce4 4800 SE?! One can argue that SE implies it is better than a NON-SE 4800. What does "SE" mean? (Special Addition?) Reminds me of the Intel 386 / 386-SX debacle.

I tend to agree with Anand here. When i see, PRO, PLUS or Ultra in front of a product (at least afa graphics hardware goes), i usually tend to think that it's just faster clock, mem. Not a different architecture (missing pipes in this case).

 

[Dave Baumann]

When i see, PRO, PLUS or Ultra in front of a product (at least afa graphics hardware goes), i usually tend to think that it's just faster clock, mem. Not a different architecture (missing pipes in this case).

Fairly ambiguous area here since it could be argued both ways. To all intents and purposes, to the consumer GF2MX was just a slower version of GF2GTS - they has nearly identical features just one was much slower than the other. For all most people knew it could have been a clocked down GF2GTS.

In this case 9500 isn't a different product as far as ATI are concerned since its the same die. To the consumer one has less fillrate than the other yet they have exactly the same features - how is this different from MX's case? How is this different from just altering clock speeds?

Fillrate may not be the biggest factor in the difference between 9500 and 9500 PRO though - its the bandwidth equality (and hence how much use they can make out of their repective fillrates) and the difference in memory sizes (hence what happens under different AA levels) thats going to cause the most confusion with these.

 

[Joe DeFuria]

When i see, PRO, PLUS or Ultra in front of a product (at least afa graphics hardware goes), i usually tend to think that it's just faster clock, mem. Not a different architecture (missing pipes in this case).

To the consumer, what is the difference between fewer pipes and a lower clock? Zero. It's simply a performance difference. When you see "SE" in front of a product, does that not mean something similar to you?

I just don't see where ATI's convention is particularly bad. If anything, they are UNDER selling the 9500 Pro, based on the comments here.

9000: DX8 part.
9500: DX9 Part, much better AA and better aniso-- surely, justifies an "increase in 500" of the number?
9500 Pro: Same architecture as 9500, only faster due to more pixel pipes. But if PRO to some people means "only faster due to clocks", then this should REALLY be something like the 9600 or 9700, right!? Architecture change!

9700: Same as 9500 Pro, only faster due to higher bus width. Should justify addition of "200" over the next lower product.

9700 Pro: Same as 9700 only faster due to higher core and memory clocks.

Again, what is something really wrong with that line-up of current generation ATI products? If you want to argue about "Pro" moniker implying only faster clocks, then I would have to believe that you should NOT be dissatisfied with the 9500 name, but think that the 9500 Pro name should be "higher."

 

[Bjorn]

In this case 9500 isn't a different product as far as ATI are concerned since its the same die. To the consumer one has less fillrate than the other yet they have exactly the same features - how is this different from MX's case? How is this different from just altering clock speeds?

Not really different. It's just that PRO has afaik always meant just faster clock, RAM. I would have prefered R 9400 or something like that instead.

I would also have prefered that they called the 9000 something else.
Especially since Ati went out with the "holier then thou" attitude with regards to the GF4 Mx and the commits the exact same "crime" themselves.

But hey, it's marketing.

 

[demalion]

You guys just confuse me sometimes. Let's see if we can agree clearly on what is bad and what is not.

Ok, launching a 8500 OEM and having it slower than a 8500 "retail" is just bad, something shameful ATi used to do. I'm sure we agree there.

Ok, launching a GF 4 MX that is just plain inferior to a GF 3 in both speed and features is just plain misleading, and people are still posting on boards (like the NwN boards, do a search for yourself) wondering why their new "GF 4" doesn't have all the fancy visuals they thought they are getting. In fact, it is worse than the above because of the inherent feature lack. Some like to excuse it as being "only" as bad as the ATi past LE problem, but reading what I just said about the result to the consumer I'm pretty sure it can be seen that it reasonably is not.

Ok, here is where things get more confusing: ATi releases the 9000...now, this isn't simple in the details because it does have newer features and all the old features (as far as the consumer is concerned) of the 8500, and sometimes it is even faster as the name implies. The problem is that in general it is slower. Now, all in all this is not good, but it is not as bad as the "unannounced" speed decrease of the "OEM" and also not as bad as the GF4 MX problem. I think that atleast the last statement can be agreed upon.

Now, to be clear, let us discuss the nVidia "4800 Ti/SE" situation Anand mentions by itself. This is arguably equal to the 9000 naming (if the GF 4 MX went poof, I'd say both companies would be in the same state as far as misleading names). You can nitpick that the 4800 SE is just plain slower than the 4600 always, and that 8x AGP isn't a "feature" for the people who buy it based on the bigger number and then proceed to use it on an AGP 4x board (and I'd agree), but in terms of product line up I'd say it is perfectly legit (and the SE isn't misleading in any real sense...the lower price will be a dead give away that the SE in this case means lower instead of higher...I do think the tag can be meant either way in all honesty).

Now is where the strange stuff (IMO) starts. First, Anand with "We can't only fault NVIDIA for misleading nomenclature though, the difference between ATI's Radeon 9500 and Radeon 9500 Pro is huge yet the difference in name implies nothing more than clock speed changes." What, excuse me? Are some of you falling for this double-speak? OK...let's see...a 9500 is slower than a 9500 Pro. Hmm...and a 9500 is faster than a 9000. Hmm...so the parallel to the 4800/4800-SE/4600/4400 is.... well... non-existant. Oh wait, "naming dishonesty" is the tie in...but...er...what naming dishonesty? Oh, 8 pipelines at a clock speed instead of 4. Hmmm...wait, isn't the only difference at all this introduces in speed, just like the name imples? Oh, wait, yes, but Anand managed to convince some of you pretty easily by unfounded implication that that is somehow dishonest. Hmm...yeah...I guess I was wrong and there is a parallel, silly me... ???!?!?!

Did some skip morning coffee or is there more to what some of you are saying?

Again, I agree it is directly comparable the 9000 naming and any distinction between the two is pure nitpicking (though to my thinking there is a difference, the difference is negligible enough to not bother discussing or even mentioning). In fact, the parallel that should be nearly perfect as ATi stopped making the 8500 and nVidia's partners will presumably stop making the Ti 4400/4600 cards. This entire 9500/Pro shtick though, I just don't get what you guys are thinking.

 

[Bjorn]

Again, what is something really wrong with that line-up of current generation ATI products? If you want to argue about "Pro" moniker implying only faster clocks, then I would have to believe that you should NOT be dissatisfied with the 9500 name, but think that the 9500 Pro name should be "higher."

Who said that something is wrong with that line-up ?
I thought we were talking about names here, not the products (which i think are great btw).

 

[Bjorn]

Ok, here is where things get more confusing: ATi releases the 9000...now, this isn't simple in the details because it does have newer features and all the old features (as far as the consumer is concerned) of the 8500, and sometimes it is even faster as the name implies. The problem is that in general it is slower. Now, all in all this is not good, but it is not as bad as the "unannounced" speed decrease of the "OEM" and also not as bad as the GF4 MX problem. I think that atleast the last statement can be agree upon.

IMO, it's even worse (R9000) then the GF4 MX problem since Ati went out and and "complained" about Nvidias naming scheme and said that they would stick to their nr = DX version so that the consumers would know what they're getting. And as soon as they got a chance, they broke their own rule.

 

[Joe DeFuria]

I thought we were talking about names here, not the products (which i think are great btw).

Well, uh, you can't talk about names and have a discussion about being "misleading" usless you also consider the products they represent.