Hardware AMD 880G Graphics Performance Review

We test AMD's latest 880G IGP graphics to see how it stacks up against the competition.

http://www.bit-tech.net/hardware/graphics/2010/05/24/amd-880g-graphics-performance-review/1

 

Nope, you spot a Christmas Edn Fenrir

 

A tad dissapointing, especially since they could've shoved a HD5xxx low end card for the IGP and have phenomal performance.

 

lol, I love the irony of this line:

"If we were feeling mean, we might point out that AMD has been using a 55nm process since 2008 (first seen on the 780G) but it's a technology that works and we'd rather AMD concentrate on shrinking its CPU and GPU dies than its chipsets."

 

Your comment about AMD "fiddling with how the game is rendered" seems a little unfounded since it doesn't appear that you tested both IGPs with the same driver revision for the screenshot you give as an example. It is entirely possible (and likely) that this is simply a driver bug.

 

Read again. AMD has to have had fiddled with the way the game worked, otherwise with both at the same driver edition the rendered scene should be the exact same. But since they're not at the same driver edition, and the output differs...

If it looks like a horse, acts like a horse, call it a horse.

 

Just a question on those chipsets when you say dual DVI, does that mean they support dual link dvi, I'm looking for a cheap ass low power integrated solution to drive my screen at 2560x1600 but am never sure if Dual DVI in these cases means two DVI ports or a dual link port.

 

Or, if it looks like a horse, and acts like a horse, it might still be a bug! I would like if they at least alluded to the possibility.

 

Eeeeeerr ....no. That doesnt rule out the possibility of a driver bug, nor does it make it unlikey. The hardware has undergone changes, the igp bios and the mb chipset for one (two). It occurs in one game, a fairly low-profile title, using the current driver, which is required and (unless your in the tin foil hat gang) strongly alludes to hardware changes requiring a software alteration to support, rather than being a benchmark result stealing driver shenanigans swizz.

The rendering errors are on this evidence a matter of good 'ol quality control not at work in the catalyst team. It could even be an issue with the mb's bios. Saying the figures are fudged for deliberate reasons is rather rash.

 

Wow I am really surprised no one picked up on this.....check out Anandtech on their "perfect HTPC card" review. Apparantly AMD does prequailifying of their drivers based on the shader count (I think starting with 10.0) and the task being required and will adjust detail levels accordingly. The driver can be forced to do the full rendering and sometimes it just "guesses" wrong. Call it a bug if you want, but it is a bug by design. I am not saying that I agree with AMD, but for most newbie types it is the simplist to get the best performance out of their hardware. I am just surpised the reviewer didn't pick up on this or at least tried to get it clairified with AMD before running the article. Shoot I even knew about, and what do I know...

 

I keep forgetting that not all people read AnandTech. That, and I need to quote sources.

I should also stop posting when sick - the replies tend to be very gruffy.

 

Hey, you had man-flu, thats all you had to say :-D

A dose of the sniffles feels like a dose of ebola for me, must be the 'y' chromosome at work.

 

I thought the largest benefit of the 880G chip was the south bridge. The previous south bridge (SB710) had performance issues with ACHI (had to use windows drivers). The performance issues were obvious with sub $100 SSD configs. So, I would get the 880G, an inexpensive SSD (indilinx based sub $100) and a WD Caviar Green for storage. This would be an inexpensive HTPC config.

I would RAID (RAID0) the SSD's also. If you tried this with 78XG you would get pokey performance at best, while using the 880, for some bench marks, the AMD chipset is faster than any Intel offering. The performance is noticeable user performance (i.e. throughput).

I recently purchased the MSI 890GX mobo upgrade from the ASUS 780G (uATX). The 780G had UI performance issues in W7 Media Center when playing back HD content (e.g TV, Blu-RAY, etc). I wouldn't go back!

 

I would get the 880 over the 78X series anytime because the pokey performance of the 78X chipset with AHCI. An inexpensive SSD config will expose the 78X chipset issues. Better $$$ spent on the 880/890 offerings.

 

This is pretty much what I've said throughout all the 8xx series reviews: unless a cheap board comes along with core unlocking, it's just not worth it over a 7xx series board. I doubt any manufacterer's will do this though, as it's a feature they've had to work hard at and they'll want to recover their investment by keeping it exclusive to their top of the range boards. That, and for all we know they may have been asked by AMD not to release it on the mainstream boards for the sake of their brand.

Either way, my upgrade is done now and I'm very happy with my i3 530. Need to play around with the OC more (all I've done so far is whack in the values from BT's review and hey-presto), but I've got a solid CPU that was cheap enough for me to splash out on a good GPU. That, and when I retire this PC it'll be perfectly suited to becomming a media PC. Now, if only I could figure out a way to make a discrete case for my old 939 system to go in the living room...

 

I used 780G for two years as my HTPC. It worked perfectly with Vista, Win 7 and Ubuntu.

The only issue with the 7xx southbridges was USB performance as far as I'm aware. I wouldnt use RAID on an AMD southbridge anyway - you're asking for trouble. You realise they just license it from Siliconimage?

 

I am am not saying my system didn't work, it just didn't perform as well as my intel system or as expected really. Also, this was a known issue. A performance issue really. Did you have sideport memory? Anyways, no problems now with my 890. SSD based 880 or 890 is still better in my opinion than 78X series with SSD.

 

The reason I chose this chipset is that it incorporates AHCI 1.2. A vast majority of MBs you see on newegg, tigerdirect, and the like are AHCI 1.1 based, which is fairly old by today's standards. AHCI 1.3 has been ratified for two years I think, and I cannot find a single MB with 1.3 support. Version 1.2, however, should (hopefully) end a lot of the headaches when installing an OS using the "AHCI" option in the BIOS for your SATA 2/3 or SSD drive.

 

Regarding this:
"AMD has literally dialed the 785G up a notch and slapped on a new number."

The only "problem" with the naming is that 785G is for unknown reason named as 7-series chipset, while it's in reality 8-series chipset.

785G is RS880, while 880G is RS880P.