AMD's 890GX Chipset - Same Graphics, Better South Bridge
by Anand Lal Shimpi on March 2, 2010 4:36 AM EST- Posted in
- CPUs
Testing the SB850’s SATA Controller
Update: New 3Gbps and 6Gbps results on the AMD 890GX here.
In my review of OCZ’s Vertex Limited Edition SSD I previewed some of the results of the world’s first SSD with native SATA 6Gbps support - Crucial’s RealSSD C300. Capable of sequential read speeds greater than 333MB/s, the C300 is the only drive in the world that can actually benefit from 6Gbps SATA at this point.
There’s just one problem. My RealSSD C300 died in the process of testing it with the 890GX. I don’t believe it was the motherboard or chipset, but right now it looks like I stumbled upon an untested usage case that put the drive in a state where it won’t even let a system POST anymore.
It’s because of situations like this that I’ve been very cautious in recommending any new SSDs. Hence my conclusion in the Vertex LE review:
“Go up another $100 and the recommendation is easily the Crucial RealSSD C300. Again, assuming that nothing horrible ever happens with the drive. I do have more faith in Crucial’s validation testing given that Micron is shipping the same drive to OEMs, but it’s still a brand new, unproven platform.”
With the bricked C300, I can’t provide any 6Gbps results on the 890GX unfortunately. I should have a new drive in about 12 hours so I’ll update here once I do get it. With the C300 out of my parts bin, I switched to a drive that could really push the limits of 3Gbps SATA - the OCZ Vertex LE.
Unfortunately, in doing so I uncovered another problem - this time with the 890GX. It’s AHCI performance is noticeably lower than Intel’s:
| Iometer 6-22-2008 Performance | 2MB Sequential Read | 2MB Sequential Write | 4KB Random Read | 4KB Random Write (4K Aligned) |
| AMD 890GX | 248 MB/s | 217.5 MB/s | 38.4 MB/s | 130.1 MB/s |
| AMD 790GX | 247.8 MB/s | 213 MB/s | 37.6 MB/s | 119.5 MB/s |
| Intel H55 | 264.9 MB/s | 247.7 MB/s | 48.6 MB/s | 180 MB/s |
AMD’s south bridge ends up delivering anywhere from 72 - 93% of the performance of Intel’s ICH. While this isn’t something that you’d necessarily see with hard drives, it is something that is evident with SSDs since they do actually push the limits of 3Gbps SATA. To make sure it wasn’t an iometer thing I also copied a 2.4GB x264 over from the boot drive (Intel X25-M G2 160GB) and still noted slower performance on AMD’s chipset. This is actually an improvement over the SB750 used in the 790GX/FX. Performance was even worse back then, particularly with writes.
And in case you’re wondering, running the SSD in Native IDE mode didn’t help either - performance was expectedly slower.
ASUS’ engineers apparently ran across something similar. They found that disabling C1E and Cool’n’Quiet boosted drive performance and recommended I try it. The results were unexpectedly higher, but not on par with Intel’s ICH performance:
| Iometer 6-22-2008 Performance | 2MB Sequential Read | 2MB Sequential Write | 4KB Random Read | 4KB Random Write (4K Aligned) |
| AMD 890GX | 248 MB/s | 217.5 MB/s | 38.4 MB/s | 130.1 MB/s |
| AMD 890GX (C1E/CnQ Disabled) | 256.4 MB/s | 234.8 MB/s | 42.3 MB/s | 135.3 MB/s |
| Intel H55 | 264.9 MB/s | 247.7 MB/s | 48.6 MB/s | 180 MB/s |
Obviously disabling important power management features isn’t a long term solution, but it does show that AMD may be able to provide a future hardware or BIOS fix for the problem.
Slower SSDs didn’t exhibit the problem. I tried the Indilinx Barefoot based Mushkin Io:
| Iometer 6-22-2008 Performance | 2MB Sequential Read | 2MB Sequential Write | 4KB Random Read | 4KB Random Write (4K Aligned) |
| AMD 890GX | 229.6 MB/s | 166.5 MB/s | 35.7 MB/s | 13.4 MB/s |
| AMD 790GX | 229.7 MB/s | 166.5 MB/s | 35.7 MB/s | 13.4 MB/s |
| Intel H55 | 236.9 MB/s | 164.4 MB/s | 36.0 MB/s | 13.4 MB/s |
It appears that the dropoff only happens in one of two cases: 1) When you’re pushing a lot of IOPS (e.g. 4KB random write tests on the OCZ Vertex LE) or 2) When you’re pushing a lot of bandwidth (e.g. 2MB sequential read tests on the OCZ Vertex LE).
As a mainstream chipset, the SATA issues don’t really matter. Unfortunately, if you are going to buy a high performance SSD then it may be an issue.
Facebook
Twitter
RSS
65 Comments
View All Comments
GullLars - Tuesday, March 2, 2010 - link
Since the SB850 also supports RAID, it would be nice to see what the bandwidth and IOPS roof is when RAIDing SSDs. The pictures say 2 GB/s bandwidth, but i'm sceptical. That would put it on par with LSI 9211-4i, only with more ports, so it's possible to reach higher bandwidth with a cheap SATA 3Gbps RAID.Dependent on scaling, you could get about 1200MB/s with x25-V's or 1500+ MB/s with x25-M's, but would the IOPS scale?
I also have a request for a re-run of your 4KB random read numbers for SB750/850 vs ICH10R, as testing at low QD on an vertex LE makes no sense when comparing avalible performance (only accesstime overhead, wich would be better meassured at QD 1). If you redo the test at QD 32, you will see numbers roughly around 150MB/s and then maybe there will be more of a difference than at ~40MB/s, especially when you consider in the additional management overhead of NCQ at higher queue depths. Even indilinx barefoot drives can do 60MB/s 4KB random read at QD 5 (wich according to your traces is within average QD in some scenarios).
sbrown23 - Tuesday, March 2, 2010 - link
I had hoped to see the Cheese Slices thing make another appearance here. Is video playback performance improved at all, or just the same as 790GX?On a side not, does anyone know how the Core i3 HD graphics does with Cheese Slices? Does it look like crap? Wondering if i3 is suitable for a Media Center/streaming media PC.
Tek92010 - Tuesday, March 2, 2010 - link
Hi, ever since the ATI purchase I've been hearing less and less about new nVidia chipsets for the AMD platform. I can understand the issues that they are having competing on the Intel side of the fence however, it really irks me to see them give up on something that they have bein doing so well for so long. In my opinion nVidia still has the best solution on the AMD side - even more so if you're running nVidia graphics.Their USB 2.0 performance, disk performance and AHCI support is up there with Intel's. Their inter-chipset communications speed is up to mark and their IGP performance is acceptable. Not to mention the general stability and maturity of their chipsets and drivers, so why then have they been so quiet on the desktop chipset side for so long?
Why haven't they updated their AMD IGP's to be on par with the excellent Geforce 9300/9400 solutions that they produced for Intel solutions? They had announced upcoming ACC support shortly after AMD started showing off what new tricks it could do with their then new SB750. How long again do we have to wait for that to be implemented?
Can someone who knows please tell me the truth about what nVidia's future chipset plans are on both the AMD and Intel fronts. I am feeling as if I will be forced to make a platform change sooner or later and I really don't want to go with AMD's current solutions until they address their relatively sub-par disk and AHCI performance.
Tek92010 - Wednesday, March 3, 2010 - link
After doing some reading last night I learned that nVidia does have an answer to AMD's ACC out called NCC (nVidia Core Calibration). Apparantly it allows for the same core unlocking and better overclocking on Phenom as AMD's ACC.http://blog.the-odyssey.co.cc/?p=179">http://blog.the-odyssey.co.cc/?p=179
Tek92010 - Wednesday, March 3, 2010 - link
After doing some research last night I discovered that nVidia had in fact released it's own version of AMD's ACC which it calls NCC (nVidia Core Calibration). It's supposed to enable better overclocking as well as core unlocking on Phenom.Paladin1211 - Wednesday, March 3, 2010 - link
Quoted:"until they address their relatively sub-par disk and AHCI performance."
It's not sub-par, mate. USB 3.0 performance is as good as Intel. And with a midrange SSD you'd have the same speed as Intel AHCI, too. Only when you slap in a high end SSD, then it would start to differ.
Given the mass majority of people don't even use a single low-end SSD let alone high-end SSD, the SB850 performs more than good enough. Enthusiasts will at least go for the i7 920 or 860, they don't deal with Phenom II 955 or SB850 either.
Sub-par disk and AHCI performance? You must be joking!
Tek92010 - Wednesday, March 3, 2010 - link
Quoted:"sub·par (sb-pär)
adj.
1. Not measuring up to traditional standards of performance, value, or production.
2. Below par in a hole, round, or game of golf."
The American Heritage® Dictionary of the English Language, Fourth Edition copyright ©2000 by Houghton Mifflin Company. Updated in 2009. Published by Houghton Mifflin Company. All rights reserved.
It's sub-par to me compared to nVidia and Intel chipsets which won't limit my nor my customer's high-end SSD performance. I was never talking about USB 3.0 performance. What qualifies you to make such a broad, generalizing statement about the processor choices of enthusiasts? Why wouldn't an enthusiast build a system with an AMD Phenom II 955 and instead opt for at least the i7 920 or 860, because you said so? The onset of SSD's have shown many people a better way to spend their system budget. Some may now opt for a cheaper and lower performing, "just good enough" processor in order to afford a good performing SSD and might end up with a better overall experience than one with an unbalanced and "mechanical hard disk drive bottlenecked" system.
It is quite posibble to build an excellent gaming system for example, based on the AMD Phenom II 955 and any of the midrange to high end GPU options available today. At high image quality settings the Phenom II 955 will surely be able to provide an acceptable and playable frame rate to many compared to it's i7 920 or 860 counterparts.
Don't forget the whole point of my post. It was really about what is going on at the nVidia camp. If nVidia were to stop making chipsets for AMD then we would all be stuck with something that is not on the same level as Intel in terms of AHCI and general disk performance. If that were to happen then it might push many system builders towards opting for Intel. Not that anything's wrong with that though.
strikeback03 - Wednesday, March 3, 2010 - link
Remember though, AMD isn't looking for just budget shoppers but ideally would like to market to the small number of big spenders too. And those big spenders won't be spending with AMD if the expensive SSD they have bought doesn't work as well.Ryun - Tuesday, March 2, 2010 - link
nVidia isn't making a LGA 1155/1366 board because they don't have a license for the QPI from Intel. Intel is legally blocking them from making chipsets for their new Core i processors (and the new Atom as well), but last I heard there was pending litigation.I don't have a concrete reason for nVidia not making AMD boards anymore, but my guess would be because a) AMD doesn't have a very good product lineup in the mobile segment and that's the one seeing the most growth at the moment. I don't think nVidia wants to invest in a platform that doesn't bring in the dough; and b) nVidia is still upset about AMD buying ATI and creating it's own in-house chipset.
That said, I agree. If I had to choose between ATI and nVidia chipsets I would choose nVidia.
Tek92010 - Tuesday, March 2, 2010 - link
Can you find out nVidia's official position on and reasoning for their lack of AMD chipset innovation and competition Anand?There was a time when the chipset market was very exciting. Are we ever going to return to the glory days of VIA vs SiS vs AMD vs nVidia vs Intel?
Also why haven't you used nVidia chipsets in your testing of AMD platforms for so long? nVidia platforms might actually make them look little better on the benchmarks than their own in-house designs, or are the AMD chipsets better than nVidia's now in your opinion?