Roadmap leaked: Nvidia GeForce GTX 580 to be 20% faster

Page 5 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

SolMiester

Diamond Member
Dec 19, 2004
5,330
17
76
Bolded part of your post is BS and you know it. Stop treating the 68xx series as the replacement for the 58xx series and actually accept that it, if anything, is a replacement for their midrange product, the 57xx.

This cannot be argued since the 69xx series will replace the 58xx series in a month or two.

If you then consider the performance increase 57xx->68xx and the same when 58xx-->69xx, you cant possibly just call it a an evergreen.

Please stop posting this particular piece of nonsense and misinformation RS. Its no good for the new forum members or new graphics card purchaser.


The part in italic/underlined: i doubt AMD can catch up with Nvidia on marketting bullshit. AMD is just not as good at it as Nvidia.

You havent been here a year and your telling a poster with years of experience to stop posting BS..LMAO!


Personal attacks are not acceptable.

Calling into question a member's credibility based on join date is an ad hominem personal attack.

RS's credibility stands on his merits, madcat's credibility stands on his...if you feel the dialogue in the thread is not civil, not cordial, or is disrespectful then the expectation is that you will report the post and let the matter be dealt with by the authorities as they see fit.

Please no more disparaging comments regarding members and their join date...the join date should be irrelevant (as should the post count), the credibility of the poster is relevant in how you view their stated position but does not give you the right to post a negative character assessment in the technical forums.

We aren't going to tell you what to think, but we are going to tell you to keep it to yourself. ;)

Moderator Idontcare
 
Last edited by a moderator:

RussianSensation

Elite Member
Sep 5, 2003
19,458
765
126
Bolded part of your post is BS and you know it. Stop treating the 68xx series as the replacement for the 58xx series and actually accept that it, if anything, is a replacement for their midrange product, the 57xx.

1) I never treated HD68xx as direct replacement for HD58xx series. This is all in your own mind. I have no idea why you keep repeating this.

Facts:

HD5750 launched at $109-129
HD5770 launced at $159, later bumped to $179.
HD6850 launched at $179
HD6870 launched at $239.

Therefore, HD6850/70 are not direct replacements for HD57xx line as you keep saying. Also, HD57xx will continue to co-exist. If you want to be correct, state that HD6850 series replace HD5830 at $180 and HD5850 at $230 and I will agree. But a $239 HD6870 has nothing to do with a $159-179 HD5770.

2) I am spouting BS to misinform new members?

More facts - HD6850/70 are Cypress-architecture chips.

"The image above maps out the major components of the Barts chip in a familiar fashion. For the most part, this is the same core GPU architecture we know from the Cypress chip behind the Radeon HD 5800 series, only scaled down slightly and tweaked in several ways." - TechReport

"Unlike what they did with the Evergreen series, AMD isn’t trying to rewrite the book on performance or push new boundaries with their “refreshed” cards. Rather, the dual engine architecture which distinguished Cypress has been generally carried over with a few modifications made along the way. Barts isn’t the focus of a fundamental architectural change in any way, shape or form." - HardwareCanucks

"The Barts GPU at the heart of these cards is the first GPU of AMD’s Northern Islands family. We’ll dive more in to its architecture later, but for now it’s easiest to call it a Cypress derivative. Contrary to the (many) early rumors, it’s still using the same VLIW5 design, cache hierarchy, and ROPs as Cypress." - Anandtech

"We expect that you'll have a tough time telling the difference between AMD's Radeon 5800- and 6800-series GPUs on this schematic. If any of this sounds familiar, it’s probably because, from a raw specifications standpoint, the new Radeon HD 6870 is essentially a Radeon HD 5830 with twice the render back-ends. " - TomsHardware

Can you please start arguying facts with facts, instead of just pointing out that my statements are "BS".
 
Last edited:

RussianSensation

Elite Member
Sep 5, 2003
19,458
765
126
What makes you think Nvidia will release first?

Sorry, I don't follow you. I only asked Groove if he thought NV could launch GTX5xx in high volume by the end of the year based on any info he may have heard. Nowhere in my post is there any mention of me saying that NV is launching first.
 

Madcatatlas

Golden Member
Feb 22, 2010
1,155
0
0
1)
Facts:

HD5750 launched at $109-129
HD5770 launced at $159, later bumped to $179.
HD6850 launched at $179
HD6870 launched at $239.

Therefore ( lol what!? my edit), HD6850/70 are not direct replacements for HD57xx line as you keep saying. Also, HD57xx will continue to co-exist. If you want to be correct, state that HD6850 series replace HD5830 at $180 and HD5850 at $230 and I will agree. But a $239 HD6870 has nothing to do with a $159-179 HD5770.

Actually, how you manage to ignore the same pricepoint is..beyond me. The 68xx series is the replacement for the 57xx cards. Your comment about the 57xx not being EoLed and thus not having been replaced is just spin, since the replacement to the 57xx ofcourse would be the 67xx, and i do hope your not going to argue ignorance of the fact that while we wanted AMD to not do change in naming conventions, they did and it was the expected (by us) 67xx cards that got repositioned as the 68xx series.


Noone here is blind. ofcourse the NI family is not a new architecture. But its different nontheless in very much the same way that the 5xxx series was different from the 4xxx series (by your interpretation, it was basicly the same exact card, just the double of everything...)
Different enough to not be called Cypress and the fact that your making me repeat this is kinda funny and a waste of my time.

Evergreen had power savings, eyefinity and about double the performance of the 4xxx series. Thats it. So lets just call Cypress/evergreen a 4xxx series derivative.
(which it is ofcourse, but do you get the point?)
 

Ares1214

Senior member
Sep 12, 2010
268
0
0
Id say a reasonable time to expect a launch, or enough stock to actually get your hands on one is february.
 

Madcatatlas

Golden Member
Feb 22, 2010
1,155
0
0
ah gsus, nvm. am i really arguing this with you RS? lol. ill leave the matter. And man am i thankful i dont have 2k posts worth of treadcrapping.
 

RussianSensation

Elite Member
Sep 5, 2003
19,458
765
126
But its different nontheless in very much the same way that the 5xxx series was different from the 4xxx series (by your interpretation, it was basicly the same exact card, just the double of everything...)

Umm...no, not even close. HD6800 series evolution to HD5800 is nothing at all like HD5800 series to HD4800 series. To claim that HD5870 was just doubled up HD4870 is incorrect. That's too simplistic.

HD5800 brought significant architectural changes compared to RV770 inside each SPs, SIMDs, TMUs and ROPs:

Shader Level (SP) changes (6 major changes in total):

1. Starting at the SPU level, AMD has added a number of new hardware instructions to the SPUs and sped up the execution of other instructions

2. Some dot products have been reduced to single-cycle computation when they were previously multi-cycle affairs

3. DirectX 11 required operations such as bit count, insert, and extract have also been added.

4. Denormal numbers were now handled at full speed.

6. Added instruction for Sum of Absolute Differences (SAD). SAD is an instruction of great importance in video encoding and computer vision due to its use in motion estimation. On the RV770, the lack of a native SAD instruction required emulating it in no less than 12 instructions.

SIMD Level Changes (4 Major changes):

1. The L1 texture cache has been improved - now capable of fetching texture data at a blistering 1TB/sec.

2. A separate L1 cache has been added to the SIMDs for computational work, this one measuring 8KB.

3. SIMDs local data share attached to each SIMD, doubled to 32KB

4. 4 L2 caches have been doubled from 64KB to 128KB. These large L2 caches can now feed data to L1 caches at 435GB/sec, up from 384GB/sec in RV770.

TMUs

1. Texture units can actually read compressed AA color buffers.

2. Texture units' interpolation speed was improved simply by them not doing interpolation. Interpolation has been moved to the SPs.

ROPs

1. Can now perform fast color clears.

2. ROPs are now responsible for handling some aspects of AMD’s re-introduced Supersampling Anti-Aliasing mode

There are other changes such as addition of a 2nd rasterizer, hull and domain shaders, and full DX11 compliance.

.
.
.
.
.
.
So you were saying that AMD just "doubled" Evergreen SPs, TMUs, ROPs into Cypress leaving the architecture unchanged.....?

On the other hand, HD6800 series is basically unchanged. All they did was just take a Cypress chip fixed MIP texture issues --> shrunk the size of the chip by throwing out ROP bottlenecked TMUs and SPs in the HD5870, and downgraded the memory controller to Redwood's, slightly retweaked the tessellation engine. You then basically have yourself a 95% Cypress chip in HD68xx. Why do you keep disagreeing with this so much? The price/performance of this chip is still outstanding. But that doesn't change the fact that this design is not of a next/new generation chip.

Look HD6800 series are great cards for the $$$, but they are just smaller Cypress chips with minor tweaks (not like HD5800 which had major tweaks from HD4800). Based on rumours that HD69xx will not be Cypress "retweak", it should be a worthy new/heavily revised architecture.
 
Last edited:

Aristotelian

Golden Member
Jan 30, 2010
1,246
11
76
Are you serious, or just being flippant?

Dead serious. You think people give more credit to posters just because they've spent longer on the forums or have more posts?

It's like making an argument that someone who is older is necessarily smarter or wiser: a basic fallacy.

On topic: Russian, I think that's all that really counts: the 6800 series cards are decent value for money, and this seems to be ushering in a wave of AMD parts that don't suffer from a huge disadvantage in crossfire when compared with their sli'd counterparts. Whether or not they are a 'big change' or a 'revision' - what does it matter? The 6900 series should be quite a bit faster than the 5850/5870 especially in tessellation (extrapolating from 6800 series performance).
 

dangerman1337

Senior member
Sep 16, 2010
333
5
81
Do any of you think that the GTX 580/GF110 will include any architechual improvements from Kepler by any chance or will it just be a improvised GF104 (or even GF100) or something else entirely?

Still if nvidia can get this chip to perform better, consume less and be smaller than the GF100/GTX48/70 i think were going to see a good start for 2011.
 

Arkadrel

Diamond Member
Oct 19, 2010
3,681
2
0
Rumors are that they cut away gpgpu power in the 580/570s... so they might be without cuda to be smaller/use less power. Again grain of salt. No idea if this is true.

Yeah the comeing 570/580 cards look to be good improvements over the 480's in gameing performance.
 

SolMiester

Diamond Member
Dec 19, 2004
5,330
17
76
Dead serious. You think people give more credit to posters just because they've spent longer on the forums or have more posts?

If you think I made that comment based solely on time spent here, I cant be bothered answering you!...Christ this forum is getting tiring!
 

Aristotelian

Golden Member
Jan 30, 2010
1,246
11
76
If you think I made that comment based solely on time spent here, I cant be bothered answering you!...Christ this forum is getting tiring!

Do you read your own posts? Here it is:

You havent been here a year and your telling a poster with years of experience to stop posting BS..LMAO!

All that your post contained was a reference to time spent posting. I'm not the only one who noticed this in your post. Paradoxically, you can't be bothered answering me but that's precisely what you tried to do: spin rhetoric as if you hadn't just called out another poster for reg date. Yeesh.
 

SolMiester

Diamond Member
Dec 19, 2004
5,330
17
76
Do you read your own posts? Here it is:



All that your post contained was a reference to time spent posting. I'm not the only one who noticed this in your post. Paradoxically, you can't be bothered answering me but that's precisely what you tried to do: spin rhetoric as if you hadn't just called out another poster for reg date. Yeesh.

Did you miss years of EXPERIENCE.....FCS, there are some pedantic fcuks on this forum. Why dont you mind your own business mate!
 

Schmide

Diamond Member
Mar 7, 2002
5,595
730
126
Did you miss years of EXPERIENCE.....FCS, there are some pedantic fcuks on this forum. Why dont you mind your own business mate!

Honestly? The comment you flipped out on wasn't even off base. The only thing off base here is your attempt to discredit someone on the basis of a limited history.

Now discrediting someone's attacks on ATI drivers based on owning an outdated used card that they had trouble with. Priceless.
 
Last edited:

Aristotelian

Golden Member
Jan 30, 2010
1,246
11
76
Did you miss years of EXPERIENCE.....FCS, there are some pedantic fcuks on this forum. Why dont you mind your own business mate!

A) I bolded the word 'experience' in my quote. Again, do you properly read the posts on this forum? Like the post from 'artonlangy' above said, what could this 'experience' possibly refer to in the context of registration dates and posting?

B) It's not pedantic to be clear with the posts here. This forum in particular has suffered a wide number of people (some compensated, others not) whose sole agenda is to distort objective data and attempt to coerce the market (the people who come to this forum looking for buying advice) into purchasing products through a biased viewpoint.

C) You are posting on a public forum, thus and so it is my business, understand?

D) I certainly am not your mate.
 
Last edited:

SolMiester

Diamond Member
Dec 19, 2004
5,330
17
76
Honestly? The comment you flipped out on wasn't even off base. The only thing off base here is your attempt to discredit someone on the basis of a limited history.

Now discrediting someone's attacks on ATI drives based on owning an outdated used card that they had trouble with. Priceless.

What?...drives?....I LOL'd at someone with not even a year posting on the forum claiming 1 of the most knowledgeable posters of GPU was spouting BS...enough already.. Why you think you have to come in a reference something from 4+ yrs ago, what a troll, FU
 

SolMiester

Diamond Member
Dec 19, 2004
5,330
17
76
A) I bolded the word 'experience' in my quote. Again, do you properly read the posts on this forum? Like the post from 'artonlangy' above said, what could this 'experience' possibly refer to in the context of registration dates and posting?

B) It's not pedantic to be clear with the posts here. This forum in particular has suffered a wide number of people (some compensated, others not) whose sole agenda is to distort objective data and attempt to coerce the market (the people who come to this forum looking for buying advice) into purchasing products through a biased viewpoint.

C) You are posting on a public forum, thus and so it is my business, understand?

D) I certainly am not your mate.

[ignore poster]