A couple of surprising things about board reviews recently

[mikeymikec]

I frequent Tech Report and Anandtech for hardware reviews mostly.

The first odd thing I noticed was that the people at Tech Report haven't done a review of an AMD board since 2011 (or even included one as part of another review, but I haven't checked every review).

The second odd thing is that Anandtech's board reviews, with regard to storage performance, seem to focus on USB performance only. While USB performance is interesting, don't people want to know how well an SSD is likely to perform on a given board (or a HDD for that matter)? My feeling is, the last time I saw benchmarks of storage performance between AMD and Intel boards, Intel was still significantly ahead of AMD. It would be interesting to know if that's still the case.

 

[Ketchup]

The second odd thing is that Anandtech's board reviews, with regard to storage performance, seem to focus on USB performance only. While USB performance is interesting, don't people want to know how well an SSD is likely to perform on a given board (or a HDD for that matter)? My feeling is, the last time I saw benchmarks of storage performance between AMD and Intel boards, Intel was still significantly ahead of AMD. It would be interesting to know if that's still the case.

Funny, I thought maybe the covered it in the initial chipset reviews and left it at that. So l looked at an initial Z87, still nothing.

http://www.anandtech.com/show/6989/...h-haswell-gigabyte-msi-asrock-and-asus-at-200

I thought I remembered some at [H], but haven't ventured over there for a while.
http://www.hardocp.com/article/2014/03/24/asus_rampage_iv_black_edition_motherboard_review/4

So they do testing, but don't compare it to anything.
http://www.hardocp.com/article/2014/03/24/asus_rampage_iv_black_edition_motherboard_review/4

So I guess you are asking the impossible!

I best if you ask that at the end of a motherboard review on AT you might get a response.

 

[mindless1]

What info are you referring to that Intel was way ahead of AMD in tests appropriate for client rather than server use?

 

[mikeymikec]

What info are you referring to that Intel was way ahead of AMD in tests appropriate for client rather than server use?

I don't have reviews to hand. My memory is of reviews that are probably 3-4 years old, however here's one that I saw before posting this thread (and it is an old review):

http://techreport.com/review/21743/...f-asus-f1a75-i-deluxe-vs-zotac-a75-itx-wifi/6

Compare the burst speed figure of the Intel system (Asus P8H67) on the last table to the rest being AMD chipset boards. A 40-50MB/sec difference on a typical HDD system would be freaking enormous (it's like comparing a 2005 HDD to a >2010 one). On an SSD system it's still worthy of an eyebrow to be raised. The other figures are fine here, they're almost identical.

As the storage system is basically the weakest link performance-wise for the average user, I'd expect both Intel and AMD to spend quite a bit of time ensuring that their chipsets/drivers can basically handle storage as well as it can be handled given the constraints of the storage device(s) in use.

It hardly is as if nothing is going on in the storage arena - HDDs have shown slow and steady improvements as usual (as far as I'm aware), and SSDs are pushing SATA 6Gbps to its limits. If there's a noticeable enough difference in USB performance to make it worthy of benchmarking/attention, then I would be very surprised if Intel and AMD's AHCI performance was neck-and-neck. Even between different boards for the same processor there's a difference.

With regard to your 'server/workstation or not' question, surely that depends? I'd say a classic server-class storage load would involved sustained I/O requests for longer periods and long disk queue lengths. A sucky desktop chipset that doesn't efficiently handle AHCI traffic is more likely to end up in the same (or similar) situation, IMO.

@ ketchup79 - I haven't ignored your response, I'll probably take your advice.

 

[mindless1]

^ The burst difference would be more important if it were enough to affect the average performance a lot but it doesn't.

The weakest link for the average user is the user, and their poor assumptions that they need the latest version of their bloated software.

 

[mikeymikec]

Which somehow invalidates the rest of the points I was making?

 

[davidthemaster3]

Wasn't there an issue with the AMD AHCI drivers being slower than the default MS AHCI drivers?

Back when I was looking to build my computer, I remember reading that Intel has the better AHCI performance...