facebook rss twitter

Review: ATI RADEON Xpress 3200 Shootout: ASUS A8R32-MVP Deluxe -v- Sapphire PURE Crossfire PC-A9RD580

by Ryszard Sommefeldt on 21 May 2006, 21:15

Tags: ATi Technologies (NYSE:AMD)

Quick Link: HEXUS.net/qafih

Add to My Vault: x

HEXUS.afterburner & HEXUS.right2reply

HEXUS.afterburner

Our Paul Dutton gives his take...

This HEXUS review sees the ASUSTeK A8R32-MVP Deluxe Mainboard rightfully awarded the coveted HEXUS.eXtreme Recommended award. However, if we had published our evaluation when ATi Technologies launched its RADEON XPRESS 3200 (RD580) core logic, on March 1, then we'd have written a very different story.

On that date, HEXUS would only have been able to conclude that, while ATi’s new chipset certainly promised a lot, the reality was that the sole production RADEON XPRESS 3200 based mainboard available to you, the HEXUS reader, was simply not fit for sale - despite what both ATi and ASUSTeK claimed. We'd also have had to say that the ATi RADEON XPRESS 3200 core logic was effectively yet another 'paper launch’.

The reason we were unable to publish back on March 1 was because the A8R32-MVP Deluxe would not correctly and reliably complete the most basic elements of the HEXUS Labs test suite even after something like a month of dedicated testing.

And remember, this was the a mainboard that had passed ATi certification and was then chosen as the launch vehicle for ATi's new RADEON XPRESS 3200 core logic chipset.

This means that the motherboard had been judged – and openly claimed – to be complete and ready for sale to the general public.

Yet, so severely fundamental were the stability issues we uncovered that, at that time, it was impossible to acquire definitive and reliable results for the mainboard's performance – nullifying, in our view, its suitability for retail sale.

For example, in his Lab notes at the time, James Smith - Manager of Performance Analysis for HEXUS - wrote:

“In FarCry (V1.33 or 1.4, it doesn't matter which), the time-demo results varied wildly. That was the case with the original BIOS (0025) shipped with the board, and subsequent point-releases, whether 0201, as shipped with the MESH system, 0307 or 0309 - the version that was supposed to fix things but didn't.

"And the variation between runs was really very wild indeed - as much as +/- 20000 FPS (Ed. yep, you read that right…) and meant that I couldn't complete testing on two of the four X1900XTX cards we had in for testing!!!!”


What was, to some degree, compounding matters was that HEXUS was also experiencing – and, please note, correctly reporting - various issues such as “crashing to desktop and poor performance when using X1900 CROSSFIRE and FRAPS”.

These we originally believed were specific to the RADEON XPRESS 3200 core logic - or should we say to the ASUS A8R32-MVP? However, as our testing ground on and on, we discovered that they weren't. They also occurred on ASUS’ A8R-MVP mainboard and this is based on ATi’s previous-generation RADEON XPRESS 200 chipset (RD480). So, it turned out, there was also a real problem with the X1900 CROSSFIRE graphics card that complicated matters.

Apart from issues with Far Cry testing, HEXUS observed and reported manifest problems when testing other gaming applications including X3: Reunion and Splinter Cell Chaos Theory.

On those particular issues, which relate to the X1900 CROSSFIRE, not the RADEON XPRESS 3200 core logic, James writes “AFAIK, to this day, these specific issues still remain unresolved”.

HEXUS first reported concerns to ATi on or around February 8, 2006 - and continued doing so for weeks afterwards. Yet the combined technological resources of ATi were insisting that the company was unable to validate our reports and had received none similar from elsewhere - presumably it meant none from other ‘publications’.

We weren't best pleased to be told that, essentially, the problem lay with HEXUS, especially when we knew that one of the myriad of ATi people assigned to resolve the issues didn’t even have a copy of Far Cry to use during testing.

The sheer volume of written and telephone correspondence relating to this debacle is shocking, and so, too, are the direct financial costs to HEXUS.

Where were ASUSTeK during all of this? Effectively dragging their feet, so ATi retrospectively claimed. And our experiences certainly laid bare serious deficiencies in the way that ATi and ASUSTeK respond to technical issues.

Of course, eventually, both ATi and, indirectly, ASUSTeK, acknowledged the reality of the issues that HEXUS had uncovered and reported, admitting that they were problems that required fixing.

On March 9, more than a week after the official launch of ATi’s RADEON XPRESS 3200 and the retail availability of the ASUS A8R32-MVP mainboard, ASUS stated it would release BIOS version 0311 for the A8R32-MVP, describing it as the "Official Qualified Version".

When ATi did acknowledge that HEXUS had indeed uncovered real issues with its RADEON XPRESS 3200 (RD580) core logic, we received a good few responses from the company, from which we offer these quotes:

"Trust me - if these guys [HEXUS] call you then it means that there is no easy way to resolve the issue - or it is a very weird one" [Meaning that if HEXUS reports a technical problem, it’s something that should be flagged at a very high level]

"Almost two weeks after you flagged the problem with us…"

"…we have finally discovered that you were correct"

"Sorry that it took us so long to confirm what you had already discovered"

"...in future, when you present a problem to us that we need to take it 100% seriously from day one and just assume that you are correct instead of trying to find flaws with your testing etc"

"Apologies for any inconvenience this may have caused"


What are the lessons to be learnt from this debacle? First, of course, that ATi Technologies and ASUSTeK each need to improve the way they respond to issues that are reported from reliable sources – and by that we don't only mean HEXUS. For instance, feedback from customers, especially early-adopters, should not be ignored.

What we've now made clear to ATi – and what other companies need to realise - is that James Smith's track record as the man who runs the HEXUS Performance Analysis operation is 100%. Each time an anomaly has been identified - and despite the occasional initial dismissive claims from manufacturers – it has turned out to be 100% correct.

But there is an even more important lesson – and it's one about who you, the would-be purchaser of PC hardware technology, can trust.

At around the time of the launch of ATi’s RADEON XPRESS 3200 chipset, at least one UK web site posted what purported to be a comprehensive review of the ASUS A8R32-MVP mainboard. This was littered with glowing praise, including this choice line “probably one of the best early BIOS motherboards that we've ever had in our labs”.

And that was posted well before ATi and ASUS fixed the problems HEXUS identified and before they'd even acknowledged them.

This surely begs the question as to how, and why, an ‘award-winning conclusion’ was posted? We also believe it calls into question the technical competence of the writer/tester and the ethics of the editor - and perhaps even the publication as a whole?

So the most important lesson to be learnt is that readers shouldn’t rely on ‘award-winning reviews’ from such bit trusted publications, no matter how much those publications might claim that their editorial and people are authoritative, competent and honest enough to responsibly guide people on prospective purchases.

To us, and others in the know, that self-evidently ain't so.

Of course what companies may choose to do, is not to seed an ethical and honest publication such as HEXUS with its products for technical editorial evaluation in a timely fashion, or at all - a tactic that’s recently not worked out well for some companies.

The track records of various other publications indicate that they are willing to try to rake in advertising money seemingly in exchange for saying nice things about existing or potential advertisers. And if they're doing that, it will be you who’ll be misguided about the true merit of products, and left paying the bill for their unethical behaviour.

Obviously it would be a very dark day indeed if you, the faithful HEXUS reader, is left, like the readers of most print-only publications, with only advertorials to read.

So, be very suspicious of companies if you know that they don’t want HEXUS reviewing what they sell. And be equally carefully of companies that don’t provide HEXUS with products in sufficient time for us to be able to properly test them before we need to publish our honest evaluation when everyone else publishes their takes – on the day those products are formally launched. Such companies will, almost certainly, also have something to hide from you.

Furthermore, we're coming to the view – honestly held – that that you need to be wary of anything that gets published elsewhere, no matter how ‘leading’, ‘independent’, ‘specialist’, ‘honest’ or ‘reliable’ the editor might claim to be.

And don't be impressed by claims of having a "review policy” “in line with industry standards" – at least, not if the ‘industry standard’ is to publish editorial so that it attracts advertisers or keeps them happy. You might as well read the advertorials which litter the newsstands.

And what of ATi’s RADEON XPRESS 3200 chipset and the ASUSTeK A8R32-MVP Deluxe mainboard?

Well, now, at last, ATi's is arguably the chipset of choice for PC enthusiasts favouring AMD Socket 939 processors – if it's implemented well. That, ATi claims, is the case for its own ‘Manta’ reference mainboard, which, it says, doesn't suffer from any of the issues HEXUS reported.

We'd very much like to have validated that claim from ATi but, despite numerous promises from several company officials, a decision seems to have been made (or a cock-up took place) that resulted in HEXUS still not receiving a ‘Manta’ reference board.

ATi blames ASUS for its initial implementation of RADEON XPRESS 3200 and the tardy fashion in which it responded to addressing the issues HEXUS uncovered. But, surely, ATi is culpable, too? The mainboard should never have passed ATi certification and been used as the ‘leader board’ to launch RADEON XPRESS 3200.

While a new face at ASUS in the UK has graciously apologised to HEXUS for the wasted time and money, ASUS have said nothing further on the matter or about its relationship with ATi.

With its initial, highly problematic, BIOS issues seemingly addressed, the ASUS A8R32-MVP Deluxe now stands as the mainboard against which all 939 mainboards based on the RADEON XPRESS 3200 will be judged. But, from our perspective, it took a worrying long time for that situation to come about.



HEXUS Right2Reply

At HEXUS we invite the vendors whose hardware we test to comment on our articles. If ASUS, Sapphire or ATI's representatives choose to do so, we'll publish their commentary verbatim.



HEXUS Forums :: 24 Comments

Login with Forum Account

Don't have an account? Register today!
just a question, with one graphics card installed, can you use eithe x16 pci-e slot or do you have to use the master (bottom) slot?

also do you know if the asus ships with the latest bios or is it still being sold with the crappy one?
Thank you hexus for keeping your ethics in place. The main reason i lurk here is due to your stance on accurate reporting. Ive been a victim of the of sites/companies dishonesty. The most frusrating thing for me was spending months trying to fix an problem that i had no chance of fixing. Strangely enough ati was one of those companies involved.( no im not an nvidia fan boy- i still buy their products when it meets my needs). The worst bit about it though was thinking “it must be me, im doing something wrong”
Blimey, talk about blowing your own Hexus trumpet. Sorry but self-aggrandisement really sticks in my throat. ;)

If Hexus had so many problems, and I dont doubt for a moment that you did, can I just ask why it took nearly 3 months for you to tell us, your faithful readers??

Surely problems of this magnitude should be highlighted to the readership asap, after the companies in question have been given reasonable time to try & fix them.

At the end of the day though it is good to hear that you can't be bought with your advertisers filthy lucre :thumbsup:b
“ means that even a lower price of just over £100”
out of interest where can you find the Sapphire board that you tested for just over £100?
The Advantage version is usually just under the £100 but the going rate for the board that you tested seems to be in the £140s (it's £165 at OcUK) …
As far as I am concerned, none of the ATI mobos have been worth a sh*t and Asus' Nvidia and ATI chipset mobos released in the past two years have been pure junk based on any number of accounts.

It's nice to see Hexus expose the dirty little secret that the Asus Xpress 3200 mobo was crap as released, as was the RD480 version. Asus has been on a downhill spiral for several years with SLI mobos not worthy of being paper weights let alone PC main boards.

In a rush-to-market mentality, virtually every single ATI RD480 and RD580 mobo has been a total POS and a rip-off of consumers. Not a single one of these boards performs properly and with Asus' circle jerk vcore BIOS voltage games, the boards aren't worth the powder to blow them to Hell. The fact that none of the ATI chipset mobos has more than one or two usable PCI slots is absurd. DFI in there wisdom decided to drop all serial or parallel ports too. Then to top off the BIOS and voltage and memory issues, the mobo makers have literally doubled the prices on these mobos without any justification.

Pigs will fly before I'd ever buy a POS mobo from Asus, Sapphire, DFI, et al. Sapphire is so lost they can't even copy the frigging reference mobo from ATI and get it right. These mobo companies have brass balls to ship a POS product that is defective in design and operation and then ignore customers and hardware review sites after the defects are documented and reported to them.

I'd suggest Asus go back to making rice cakes as they haven't a frigging clue about how to produce PC main boards.