ATI Radeon HD 5670: DirectX 11 For Under $99

[cleeve]

[citation][nom]reasonablevoice[/nom]12-20% is actually pretty signifacant for cards in a similar price bracket. [/citation]

It's not a 12-20% performance difference, more like a a 0%-20% performance difference, depending on the game.

So why is a 0%-20% performance difference 'significant', but the guaranteed 20% price difference is irrelevant to you?

Tsk, Tsk! Not quite the reasonable kind of comment I'd expect from a fellow with your handle.

 

[cleeve]

[citation][nom]casecutter[/nom]Don Woligroski, that was the most jaundice, craftily distorted article I’ve read lately.[/citation]

Funny! That's how I feel about your post.

Your re-hashed complaint has been thoroughly addressed on page 3 of the comments. If it's not enough to quell your conspiracy theory, well... you have my pity.

 

[Casecutter]

"Objective journalism and an opinion column are about as similar as the Bible and Playboy magazine."
Walter Cronkite
What do you supply?

On page 3 you said,

If this (5670 review) was biased, I'd be extolling the virtues of the GT 240

, although the silence speaks volumes. Most everyone considers the 5670 to weigh in against Nvidias’ newest and most direct contender (GT240), which has the distinction of being the "card that doesn't matter". They're essentially for plug-and-play upgrade of stock OEM systems that the average novice can undertake without a PSU, and not those higher powered gaming cards. To even be able to say such a card can bite at the heals of last generation stuff (the elderly veterans of a really ugly price war) is commendable.

The bad part is yes, the group (4670, GT240's and now the 5760) is saturated and we need a price need and will settle out. So, on that point we agree… 5670 1Gb need to quickly settle out to $85-99 and 512Mb down to $70. then they'll have their rightful place in the mix. As to the GT240 "it won't matter".

 

[kettu]

[citation][nom]Cleeve[/nom]What I admit is that I was poking fun at your shared obsession with insignificant vernacular.[/citation]

What you call "insignificant vernacular" is actually erroneous interpretation of data. But, whatever.

 

[Guest]

Great article. Good to know that ATI has improved in reducing
power-consumption/temperature @ idle. I would definitely recommend this product to my price-conscious friends [who wanted something great for playing L4D, CallOFDuty, CompanyOfHeroes].

Good selection of cards for comparison.

 

[cleeve]

[citation][nom]casecutter[/nom]...the silence speaks volumes. Most everyone considers the 5670 to weigh in against Nvidias’ newest and most direct contender (GT240), which has the distinction of being the "card that doesn't matter". [/citation]

My silence speaks to the relevancy of the GT 240.
If 'most everyone' wants to concentrate their attention on the 'card that doesn't matter', then I choose not to waste my time following their lead.

As for what I supply, my brand of objective journalism compels me to focus on the 9800 GT, 4850, and GTS 250 as the price/performance alternatives.

I can only assume you agree with my recommendations regarding the 4850/GTS 250 as better price/performance alternatives. You also seem to agree with my conclusion that for 'folks who don't want to upgrade their power supply, the Radeon HD 5670 does offer the fastest reference card performance you'll find without connecting a dedicated power cable.'

Yet you call the article 'jaundiced' and 'craftily distorted'. But you agree with my conclusion almost verbatim. That doesn't make a lot of sense.

Objectivity is often in the eye of the beholder unfortunately, but the pursuit of the ideal is key. To be frank, it does look to me like you might have read this article with a preconceived notion of bias based on an article I wrote about the GT 240 -- an article that I have admitted is flawed due to inaccurate pricing assumptions.

I suspect if you hadn't seen my GT 240 article you'd have been less critical. I have explained myself, apologized for it, and most importantly I have learned from my mistake so that I do not repeat it.

 

[cleeve]

[citation][nom]kettu[/nom]What you call "insignificant vernacular" is actually erroneous interpretation of data. [/citation]

And what you call "erroneous interpretation of data" is actually "bickering over the minutia".

See? It's the great circle of life, coming back to bite you in the arse.

 

[noob2222]

[citation][nom]Cleeve[/nom]It's not a 12-20% performance difference, more like a a 0%-20% performance difference, depending on the game.[/cititaion]

And apparently on how your testing it, Anandtech got 12% faster than the 9600 in crysis, yet your review put the 9600 being faster.
http://www.anandtech.com/video/showdoc.aspx?i=3720&p=5

[citation][nom]Cleeve[/nom] So why is a 0%-20% performance difference 'significant', but the guaranteed 20% price difference is irrelevant to you? Tsk, Tsk! Not quite the reasonable kind of comment I'd expect from a fellow with your handle.[/citation]
You say the 5670 = 9600, yet its 0-20# faster (for the sake of arguing), however you say the 9800 is faster, yet the perfomance difference is only -2% to 5%.

Here is another clue

So where does that leave the 5670? The 5670 does surprisingly well against the 9800 GT. It wins in some cases, trails very slightly in a few more, and then outright loses only in games where the 5670 is already playable up to 1920x1200. From a performance standpoint I think the 9800 GT is ahead, but it’s not enough to matter; meanwhile the “green” 9800 GT shortens the gap even more, and it still is over 10W hotter than the 5670. The 5670 is a good enough replacement for the 9800 GT in that respect, plus it has support for DX11, Eyefinity, and 3D Blu-Ray when that launches later this year.

http://www.anandtech.com/video/showdoc.aspx?i=3720&p=14

 

[bc0203]

I think the HTPC point is valid. The triple combo of HDCP, DTS-MA and Dolby Digital HD on the card, with 1080p support, will make this card a good value once the "new" pricing has worn off and it hits a $80-585 price point.

 

[cleeve]

[citation][nom]noob2222[/nom]...however you say the 9800 is faster, yet the perfomance difference is only -2% to 5%.[/citation]

I think I can see why you're getting worked up over the minutia here. You're focused entirely on absolute performance, and I'm focused on performance per dollar.

The 9800 GT be found for ~10% cheaper than the 5670 despite performing better. To me that's more significant than the absolute margin of performance difference. IMHO, both factors have to be taken into account for a meaningful analysis. It's not about absolute performance, it's about the performance/cost ratio... if it wasn't, every card in the world would suck except the 5970.

You want to me say the 5670 is better than the 9600 GT? The data already says that. Sure it is. It just isn't a better for the money spent.

But the 9600 GT often equals 5670 performance for $20 less, and the 9800 GT beats it for ~$10 less the majority of the time. You can argue about a percent here or there all you want, but it doesn't change the fact that the 5670 needs to be a little cheaper than the 9800 GT to be the fabulous buy it can be.

As far as Anand: A great bunch of guys, but I have to disagree with the approach they used to benchmark the 5670. Instead of tailoring the 1680x1050 and 1920x1200 settings to the card, they benched it at the same settings they used for higher end cards - probably to make use of previous benchmark data they had onhand.

This would save them a lot of work, and I can't blame them for that - AMD didn't give us a huge amount of time with the card - but this results in benchmarks that have little real-world relevance. Who is going to play at 1680x1050 settings that will drop this card down to 23 FPS in Far Cry 2? 12 FPS in Crysis: Warhead? These settings and results don't show us how a real human being would use the card, and therefore don't do the 5670 justice.

 

[bc0203]

I think the HTPC point is valid. There plenty of people out there who aren't gamers with who have HTPCs based on Intel G4x/P4x chipsets who will want a cheap way to play back Blu-ray discs with DTS-MA and/or Dolby Digital HD content without having to upgrade their whole system. This fits the bill perfectly, if the price point can be lowered to under $80, which I predict will happen once the "newness factor wears off.

 

[noob2222]

[citation][nom]Cleeve[/nom]I think I can see why you're getting worked up over the minutia here. You're focused entirely on absolute performance, and I'm focused on performance per dollar.[/citation]
ya, whaterver. Now your going to change the words in your conclusion that has no mention of performance per dollar, yet you still want it to equal the price of a card thats up to 20% slower.

Exact same could have been said about Nvidia's 240 review, but instead it was accepted as perfect since its Nvidia, but hell bent if ATI can try to make an extra $5 on thier new cards. The 240 is slower than the 9600 and costs the same $99. Consider the article as what it was wrote for, sub $100. Here is the conclusion for Nvidia (you have to ignore the fact that it released at $110.)

But none of these issues are show-stoppers today. The point is that Nvidia now has a cost-effective part that it can leverage to not only compete with the existing Radeon HD 4670 when paired up to DDR3, but replace the GeForce 9600 GT when it's armed with GDDR5. Though at-launch pricing is usually high, Nvidia will finally have the flexibility to compete at the entry-level once production has ramped up, and we will undoubtedly see that happen with the GeForce GT 240.

So it would appear to be quite fine for Nvidia to release a sub $100 card that performs less than thier $80 cards, but not ok for ATI to do the same.

 

[cleeve]

ya, whaterver. Now your going to change the words in your conclusion that has no mention of performance per dollar, yet you still want it to equal the price of a card thats up to 20% slower.

Haven't changed words in my conclusion, nor do I 'want' anything from any of these graphics cards, except to perform well for the price.

You're stretching, noob. You don't even have any arguments anymore, you're just spraying sour grapes.

 

[noob2222]

Your ignoring everything fanboy. Let me ask you a straight question. Why do you think this card should be priced $80? Because of crysis?

Here is a clue:
Add the totals of all the benches and average it out.
5670=67.5fps
9800=70.2fps
9600=60.4fps

Show me how the 5670 = $80 = 9600gt, thats clearly stated in your conclusion.

 

[cleeve]

[citation][nom]noob2222[/nom]Your ignoring everything fanboy.[/citation]

Heheh. I love the smell of irony in the morning. Smells like victory.

[citation][nom]noob2222[/nom]Let me ask you a straight question. Why do you think this card should be priced $80?[/citation]

Because its performance does not justify the $20 spread over the $80 9600 GT.

Your turn. Why do you think the 5670 is worth $100 to a gamer when the 4850 can be purchased at the same price with vastly superior performance?

Straight answer now! No more sour grapes...

 

[noob2222]

Read my first post since you ignored it also. I said $90 would be ideal, however $100 isn't out of the question considering the extra features as well as lower power consumption.

And seeing as the performance is between the $80 9600 and closer to the $95 9800, logically and ignoring any features or power, $90 is still a better price.

 

[kettu]

[citation][nom]Cleeve[/nom]And what you call "erroneous interpretation of data" is actually "bickering over the minutia".[/citation]

I averaged the results of highest resolution and found the following:
9800 is on average ~8% faster than 5670
9600 is on average ~10% slower than 5670

So it looks like "erroneous interpretation of data" is what we're actually talking about here.

 

[kettu]

Forgot to add:
Since you probably have those results in a convenient form (excel?) you can easily check wether I made mistakes or not.

 

[cleeve]

[citation][nom]kettu[/nom]So it looks like "erroneous interpretation of data" is what we're actually talking about here. [/citation]

One man's erroneous data is another man's bickering over minutia...

What surprises me is that at this point in the game you actually think I'm interested in engaging you in your unhealthy obsession with dissecting my adjective.

 

[cleeve]

[citation][nom]noob2222[/nom]I said $90 would be ideal, however $100 isn't out of the question considering the extra features as well as lower power consumption.[/citation]

So the basis for your forum rage is that you're appalled by a $10 difference in our appraisals?

To me $90 is still to close to the 4850's $100. I cant imagine recommending to save $10 in exchange for the ability to play games at higher detail at 1680x1050 or even 1920x1200.

Maybe you can, that's fine. There's room in this world for more than one opinion, I'm happy to agree to disagree with you over a $10 perception in value.

 

[kettu]

[citation][nom]Cleeve[/nom]One man's erroneous data is another man's bickering over minutia...What surprises me is that at this point in the game you actually think I'm interested in engaging you in your unhealthy obsession with dissecting my adjective.[/citation]

You keep replying with inanities, though. So I guess you do have some reason to engage me, you just don't do it very well.

 

[cleeve]

[citation][nom]kettu[/nom]You keep replying with inanities, though. So I guess you do have some reason to engage me, you just don't do it very well.[/citation]

I guess I wouldn't be doing very well if my motivation was to bicker over irrelevant details.

However, if my motivation was to poke fun at your unhealthy obsession, I think my success speaks for itself.

 

[mayne92]

[citation][nom]frederico[/nom]...Seriously, what is with all the bias on this site toward Intel and nVidia, its like Fox news 'fair and balanced' reviews.[/citation]
Honestly where is there any "unbiased" information anywhere? People tend to conform with their favorite sides...which is typically...natural. Nothing against the article at all or Tom's for that matter, but people seem to come here for their one-stop-news-place when they should be GATHERING information from several sources and developing their own conclusions. Maybe you people are just fucking lazy?...or like to bitch a alot? BTW, Fox just likes to concentrate on "proving" the other new sources "wrong"...IMO I can't blame them either...but sometimes they take it a little too far.

 

[hannibal]

Honestly where is there any "unbiased" information anywhere? ...they should be GATHERING information from several sources and developing their own conclusions.

Good points. It is allways better to read many articles, but it is allso good to have some reference sites. And when chosing to those reference sites we come to area where things like "unbiased" flyes out of window!

What most sites seems to agree is that these new cards from ATI and Nvidia are too expensive when compared to older models. We allso knows that these new cards are (should be) cheaper to produce to Nvidia and ATI than those older models, so we can expect either price cuts when they can produce these enough or they drop out their more expensive older models (and that way making these the only alternative to consumers) or both.
They also seems to eat less power (surprice), so they are now viable alternative to those who need GPU that eat less power. We can only hope that this is not the starting of the trend, when PC-are getting more expensive again...

 

[mayne92]

[citation][nom]Schip[/nom]Small typo on the last page: "..., the Radeon HD 4670 does offer the fastest reference card performance you'll find without connecting a dedicated power cable."That should probably say '5670'.Coming from one who read the article carefully, I believe the article was well written. Every statement was quite fair, and quite justifiable. A very worthwhile article. Thanks![/citation]
I agree. Good article! I also was able to take the information and make my own conclusions as well...does anybody need a tutorial? lol.