The Card that Killed the Dragon

[Computer Ed]

Below is an exceprt from my blog of an entry that I have posted concerning a story in tech news that seems to not be getting coverage. Chris Hook from ATI told me he thought it deserved to be covered and wondered why it was not.

Now we are looking at the next generation of video cards from ATI. The reviewers all agree the 5850 is redefining the mainstream with the power it is bringing at a mainstream price. Yet as we look at those reviews and take a a deeper look, something is different. First with this release, unlike past releases we are not seeing any mantra of experience over benchmarks. Conspicuously absent as well is any mention of how this card impacts Dragon or for that matter if this moves the platform concept to a whole new level. AMD seems in fact kind of silent on a front they have been screaming about for a few years now.

The link is below for the entire story is below..



Link removed as it violates the self-promotion rules of this forum.

Video Mod BFG10K.

 

[alcoholbob]

AMD knows they aren't going to win in the enthusiast market with their CPUs, so the last thing they want people to do is think of "AMD the CPU" when they see the ATI logo with the new HD5xxx series, now that they are in enthusiast territory. Brand separation is a good idea in this case and it's smart marketing.

 

[Idontcare]

self-linking/promoting is against the rules here

 

[Corporate Thug]

it's clear at this point that i7 is the fastest gaming cpu...doesn't it only make sense that i7 be used as the platform?

 

[Corporate Thug]

Originally posted by: Idontcare
self-linking/promoting is against the rules here

oh, and this.

 

[lopri]

Umm.. maybe because that strategy didn't work with HD 3000 series and HD 4000 series? Plus, even when AMD pushed that strategy review sites ignored it for the most part anyway. For obvious reasons.

 

[manimal]

Well if they kill the dragon they can restart the game on more diffiicult settings and do it all over again...



Maybe they dropped the marketing spin about dragon because few really cared-except fanboy nabs- dont get me wrong I own 5 amd procs now and 5 ati cards from the last generation and I still didnt give a meh about their so called platform. IF and thats a big IF scaling was good with quad crossfire I would care alot more but it doesnt so......

 

[Idontcare]

Originally posted by: lopri
Umm.. maybe because that strategy didn't work with HD 3000 series and HD 4000 series? Plus, even when AMD pushed that strategy review sites ignored it for the most part anyway. For obvious reasons.

Exactly...should be dubbed "the card that stopped beating a dead horse".

Originally posted by: Computer Ed
Chris Hook from ATI told me he thought it deserved to be covered and wondered why it was not.

Ask Chris to ask his superiors. I'm sure if AMD really wanted it to become a media story then AMD would have included the necessary talking points regarding it in their presskit for the HD5870/5850 reviews. Wouldn't they?

 

[v8envy]

Correct me if I'm wrong, but doesn't Phenom work better with NV cards than ATI cards at the moment? I'm sure the weird performance issues with the 4 (and 5?) series hardware on Phenoms will be dealt with at some point. But until then, i7 + ATI and AMD + Nvidia are the best gaming combinations.

 

[manimal]

Originally posted by: v8envy
Correct me if I'm wrong, but doesn't Phenom work better with NV cards than ATI cards at the moment? I'm sure the weird performance issues with the 4 (and 5?) series hardware on Phenoms will be dealt with at some point. But until then, i7 + ATI and AMD + Nvidia are the best gaming combinations.

How so? I have been under a rock lately -work + wife pregnant with triplets- What have I missed?

 

[dguy6789]

Originally posted by: v8envy
Correct me if I'm wrong, but doesn't Phenom work better with NV cards than ATI cards at the moment? I'm sure the weird performance issues with the 4 (and 5?) series hardware on Phenoms will be dealt with at some point. But until then, i7 + ATI and AMD + Nvidia are the best gaming combinations.

Definitely wrong. Intel processors are better across the board for gaming. i7, i5, Core 2 Quad, all better than Phenom II and lower AMD processors.

 

[OVerLoRDI]

The only thing AMD has going for it in the CPU market is price/performance ratio. 100 dollar quad cores is legit.

 

[Zstream]

Originally posted by: Idontcare
self-linking/promoting is against the rules here

Then tell ABT to stop...

 

[toyota]

Originally posted by: dguy6789

Originally posted by: v8envy
Correct me if I'm wrong, but doesn't Phenom work better with NV cards than ATI cards at the moment? I'm sure the weird performance issues with the 4 (and 5?) series hardware on Phenoms will be dealt with at some point. But until then, i7 + ATI and AMD + Nvidia are the best gaming combinations.

Definitely wrong. Intel processors are better across the board for gaming. i7, i5, Core 2 Quad, all better than Phenom II and lower AMD processors.

not wrong. its certainly weird but surely you should have read about it here on Anandtech since I believe its been mentioned in more than one article. http://www.anandtech.com/mb/showdoc.aspx?i=3639&p=3

also in Far Cry 2, bit-tech got way better performance out of AMD Phenom 2 and Core 2 than i7 when using an nvidia card. http://www.bit-tech.net/hardwa...4-940-and-920-review/7

 

[mmnno]

Originally posted by: v8envy
Correct me if I'm wrong, but doesn't Phenom work better with NV cards than ATI cards at the moment? I'm sure the weird performance issues with the 4 (and 5?) series hardware on Phenoms will be dealt with at some point. But until then, i7 + ATI and AMD + Nvidia are the best gaming combinations.

You've got it backwards. The issue is not with 4 and 5 series hardware and Phenoms; ATi cards work perfectly fine on AMD chipsets. The issue is with nVidia cards on intel chipsets.

 

[OCGuy]

Ban

 

[Lonyo]

Actually this thread raises an important point that places like AT should take note of.
They run one card on one CPU and test it against other cards on the same CPU.
In CPU tests they take one card and put it with multiple CPUs.

Has any site done a review of the HD58xx series which puts a Phenom II 965 against an i5 750 and i7 975? (Or similar).
I mean sure, we expect the i7 to beat the PII, and the i5 will probably be around the same level, but where are the platform tests? AT has already shown that there is some variability and it's not always cut and dry, and in many gaming situations the CPUs seem to be close (e.g. 5% difference). With less of a GPU limit it would be interesting to see what sort of performance comparisons can be made between the CPUs.

Who cares if an i7 975 is 8% faster on average when a PII is 20% of the cost? People buying a $260 graphics card won't care, they will want the best value for money to extract maximum performance from their graphics card.

 

[Lonyo]

Originally posted by: Corporate Thug
it's clear at this point that i7 is the fastest gaming cpu...doesn't it only make sense that i7 be used as the platform?

In Triple/Quad SLI/Crossfire, yes it makes absolute sense. If you have the money for that many cards, that big a PSU, you can afford an expensive CPU.
In a review of a single $260 graphics card? Hell no. Not everyone will want to spend $400 on an i7 + mobo + RAM when they could spend more in the region of $250 or even less (Athlon X4's, low end Core 2 Quads) and get most of that performance and put it with a $260 graphics card and buy the whole core of a system for $500.

 

[BFG10K]

Originally posted by: Zstream

Then tell ABT to stop...

To my knowledge nobody affiliated with ABT posts links outside their sigs.

If you see any examples of this happening, please forward them to me.

Having a link in your signature is allowed under the current rules of this forum. Having a link in your post (like the OP) is not.

Video Mod BFG10K.

 

[Computer Ed]

My apologies for the perople that think this was a spam effort. I was really only interested in people reading and discussing the actual article I posted the excerpt from.

I am looking forward to following up with AMD and use the responses from the forums I have posted in as things to discuss with them.

The reason this whole scenario stood out to me is easy to understand if you look at the recent history of ATI releases. Each time they have pushed and stressed the platform position and this time they seemed to completely ignore it.

Additionally do you really thin all these websites together without exception decided that they would ignore such an OBVIOUS review point as seeing how this fits into the Dragon platform? I can believe MOST doing this only on Intel but ALL?

I posted the article to get this being discussed on forums, not my site, so thank you all for opening a discussion on this.

 

[dguy6789]

Originally posted by: toyota

Originally posted by: dguy6789

Originally posted by: v8envy
Correct me if I'm wrong, but doesn't Phenom work better with NV cards than ATI cards at the moment? I'm sure the weird performance issues with the 4 (and 5?) series hardware on Phenoms will be dealt with at some point. But until then, i7 + ATI and AMD + Nvidia are the best gaming combinations.

Definitely wrong. Intel processors are better across the board for gaming. i7, i5, Core 2 Quad, all better than Phenom II and lower AMD processors.

not wrong. its certainly weird but surely you should have read about it here on Anandtech since I believe its been mentioned in more than one article. http://www.anandtech.com/mb/showdoc.aspx?i=3639&p=3

also in Far Cry 2, bit-tech got way better performance out of AMD Phenom 2 and Core 2 than i7 when using an nvidia card. http://www.bit-tech.net/hardwa...4-940-and-920-review/7

That's extremely strange and almost certainly a driver or bios issue of some kind.(That the Nvidia cards are performing better on the Phenom II) With some very minor exceptions, pretty much every benchmark out there shows the i5, i7, and C2Q being superior processors clock for clock to the Phenom II in just about everything. Considering that, it definitely seems that this anomaly is a driver or bios issue or something of that nature.

 

[Idontcare]

Originally posted by: Computer Ed
I am looking forward to following up with AMD and use the responses from the forums I have posted in as things to discuss with them.

I'm confused, in your OP you said the progenitor of your writing a blog peice on the topic was an AMD employee...why wouldn't you just have the discussion with him?

What does "AMD" have to say that this particular AMD employee does not?

Originally posted by: Computer Ed
Additionally do you really thin all these websites together without exception decided that they would ignore such an OBVIOUS review point as seeing how this fits into the Dragon platform? I can believe MOST doing this only on Intel but ALL?

I don't quite follow you here...what is it that you are trying to say/imply?

That AMD instructed the reviewers to use Intel rigs and intentionally told them to avoid using AMD rigs as their benchmark platform for HD5870/5850 reviews?

Given Anandtech's history of "outing" such PR requirements from past AMD marketing events I would be quite surprised if this did transpire but Anandtech elected to not mention it this go around.

This is info I would want to know, because if it is true then it's omission is relevant to me at a personal level.

 

[Wreckage]

Well when most sites run CPU tests they usually use a NVIDIA card. Even with an AMD CPU.
http://www.anandtech.com/cpuch...ts/showdoc.aspx?i=3638

When testing a video card you want the fastest CPU to eliminate that as a bottleneck and vice versa.

 

[Computer Ed]

Originally posted by: IdontcareI'm confused, in your OP you said the progenitor of your writing a blog peice on the topic was an AMD employee...why wouldn't you just have the discussion with him?

What does "AMD" have to say that this particular AMD employee does not?

Sorry for any confusion. I will attempt to speak again to Chris Hook for a followup but you never know who AMD will want you to talk to when you approach them. So with this in mind I generically refer to the company as a whole instead of an individual.

I don't quite follow you here...what is it that you are trying to say/imply?

That AMD instructed the reviewers to use Intel rigs and intentionally told them to avoid using AMD rigs as their benchmark platform for HD5870/5850 reviews?

Implying is a strong word, I am opening a dorr for discussion because this is very odd behavior. I mean you have AMD preaching a mantgra for so long of platform and then a bunch of sites, all with access to the AMD platform review two different AMD video cards and not ONE decides to even add the platform as part of the review? Your telling me that does not seem strange to you?

This is info I would want to know, because if it is true then it's omission is relevant to me at a personal level.

This is why I posted the artile on my blog in the first place, to drive discussion of a topic that no one seemed to even notice, including AMD. That in itself seems odd, surely someone at MAD must have noticed this but when Chirs Hook and I spoke he seemed genuinely suprised.

Wreckage your point is valid but I want to put an emphisis on the one word that makes my point MOST. Not all, not always but MOST.

 

[lavaheadache]

Originally posted by: Wreckage
Well when most sites run CPU tests they usually use a NVIDIA card. Even with an AMD CPU.
http://www.anandtech.com/cpuch...ts/showdoc.aspx?i=3638

When testing a video card you want the fastest CPU to eliminate that as a bottleneck and vice versa.

what was the point of this statement? How can you even remotely prove this? Why am I even concerned by this? Does this statement somehow reflect my probable correctness regarding your apparent perpetual quest to aquire such status as an Nvidia focus group member? Who knows....