Thanks Raja for the excellent review - I've held back on purchasing a motherboard awaiting 'Part 2'. I'm looking for a relatively low power consumption board for a windows home server setup.
I was actually pretty amazed by the Gigabyte idle power results - as equal to the the MSIboard (44 watt).
For much of the time the CPU will be used minimaly - only rising for the occasional media decoding task - so for me the Gigabyte boatrd seemed ideal - partcularly since for some HD media sata3 speeds might come in useful.
But the section on power consumption concludes
"Best overall power consumption figures belong to MSI, while boards laden with USB 3.0 features and SATA 6G trail in both idle and load situations."
The Gigabyte board (H55-USB3), quite remarkably doesn't appear to trail at all on idle.... am I missing something here?.
what did you mean by:
"particularly since for some HD media sata3 speeds might come in useful."
The board doesn't have SATA3 does it? Only USB3.
And what HD media would benefit from SATA3 speeds?
There isn't any HD media I'm aware of that'd come close to saturating sata2.
Or am I misunderstanding what your were trying to say?
No misunderstanding - I probably need to say a little more.
I currently one one Sagetv media extender off the server. But am about to install two more. The system should ideally allow for 3 HD videos to be streamed off the same HD at the same time.
I haven't done the calaculation - but having the option for SATA3 seemed like a good idea.
I've also re-read the review a note that this board has "Gigabyte SATA 2 chip: 1 x IDE, 2 x SATA 3Gb/s (RAID 0, 1 and JBOD)"
and NOT 2 *SATA3 g/bs ... my misreading,sorry for the confusion.
Rod
Hi,
Thanks for the feedback. The comment was made considering all three scenarios (Idle/full load/video playback), the MSI board came out on top in all three, hence the comment 'best overall power consumption'.
I would love to see an Anand review of the Intel DH57JG and Zotac H55ITX-A-E motherboards. I am looking to put one of them (probably the Intel) along with an i5-650 into a Lian Li PC-Q07 case. I love to build powerful systems for others but stick to small and simple for myself.
First, thank you for the detailed information on fan control. I've made a few comments about that in the past and it's nice to see that category now seems to be part of a standard Atech mobo review.
I'm not in the market for a Clarkdale platform, but if I was, this article would be very useful.
A lot of people are talking about new motherboards being released before they are ready. And for good reason. Why bother? Why not go and buy a well known and well established motherboard like the GIGABYTE GA-G31M-ES2L for $45 and drop in a E5200 and a quiet GT 220 or 5450? Overclock it to a modest 3 GHz and it will surely smoke this H55/H57 garbage in all the gaming benchmarks, for a LOT cheaper.
I do not understand the value in this entire product line. Why do you not compare these with the option I just mentioned? I dont care about how intel wastes their monopoly advantage. If hardware from a year ago is cheaper and better than this crap they are shoveling out now, then it is your responsibility to tell us that.
Point taken, but as readers, isn't it great that we don't have to buy all the newest, most expensive hardware to find out how it runs? And the problems with it? We can just read articles like this one.
The Anandtech folks *do* write articles showing budget parts that offer exceptional value (via overclocking or unlocking cores). This just isn't one of those articles.
Thank you for pushing Gigabyte on the AHCI issue! Can you ask them about their 790-series boards, too? I'm frustrated with the lack of AHCI support on my new GA-790-XTA-UD4 motherboard. There's a 30-45 second delay in initializing SATA hard disks when returning from sleep mode. This causes Windows 7 to blue screen. The only fix is to revert to IDE mode for all drives. I wasted 2 days trouble-shooting this only to find out its an unaddressed problem with AHCI support on this motherboard (and many others).
It's funny because the hotfix definitely prevents total operating system failure (i.e. BSOD). However, Windows takes up to a full minute to become responsive when resuming from sleep mode. Presumably the OS is waiting as long as it needs to for the SATA boot drive to become responsive again.
I will continue running in Native IDE mode for now, since losing 1% system performance is less irritating than waiting forever for my system to become responsive.
*** It would be nice if AMD or Gigabyte addressed the true problem, but I won't hold my breath. I still haven't heard back from their customer support about this problem and it's been a few days. Not even a "we'll get back to you soon" message. Nada.
Great article Raja, I also appreciate the detail of the good and bad!
I have a question for you or anyone else who might know - you mentioned ASUS dropping the MSRP of their H55 board at the start of the article... Where can I look up what the different vendors MSRP's are? Intel and AMD have made it quite easy to find out the 1K unit price of their CPUs on their own sites, but I haven't seen anything similar for motherboard vendors. Is there an authoritative, frequently-updated source?
Why does the MSRP really matter? Unless you are looking to buy in bulk, I don't really see what the benefit would be. For instance the Intel 80gig SSD has been inflated for months now. It's still a great product, and I don't know why someone wouldn't buy something if the price was higher than MSRP. Many electronic products sell below MSRP, and IMO that does't make them any better or worse.
More importantly than MSRP in my opinion is shopping around to get the best price at any given point. 15-30min of googling around to various E-tailers will give you a good ballpark of the price.
Unfortunately, I haven't seen or heard of any such resource from any of the sub-vendors - I usually have to fire off emails to ask. I agree with you though, it would be nice to see the figures posted on vendor websites.
Samsung, a respected brand keeps introducing mobile phones that match today’s speed and technology. Samsung has gone great lengths this time as it has received impressive reviews for its new Galaxy S 4G which happens to be the first 4G Phone with 21+ Mbps speeds.
(ASUS is using a 188 amp capable VRM vs a guessed 110~120amp VRM on the Gigabyte H55/H57 boards, so we'd pick the ASUS boards for Lynnfield CPU overclocking)
Looks like the units used to express the capabilities of the VRM's is incorrect, otherwise I am going to need a really large PSU to run them....
I don't understand these Farcry 2 benchmark results.
How is this score against the i5 750 being calculated? Did you test all these boards with that processor then average it? :S
If the first result is the 661@stock and the second the 661@4Ghz, why is the comparison to the i5 750 number not the same in both charts? What changed for the 750? The graphics settings appear to be the same.
The 750 was benched on the EVGA P55 FTW board. Same drives, same drivers, same GPU, same patch version of FC2. Not sure why Lynn (4core) does worse in this bench.
If the first result is the 661@stock and the second the 661@4Ghz, why is the comparison to the i5 750 number not the same in both charts? What changed for the 750? The graphics settings appear to be the same.
The 750 was OC'd to 4GHz in the second chart, which is why the figure is 5FPS higher than the stock number of 68fps.
I can update the graph in about a day with the 750 running in one of the H55 boards if you like - got the Intel JG on the test rig at present and it won't do 4G's with the 750 I suspect..lol
Okay, USB 2 transfers a gigbyte in 36 seconds. This is consistent with the rated spec of 480 Mbit/s. Isn't USB 3 supposed to be 10x faster? What's causing the limitation? Is it the implementation by Asus and Gigabyte (like attaching the USB 3 to too narrow of a PCI-e lane)? Is it simply the write speed limitation of the USB HDD used? If the latter, why not write to some flash memory to see if the results improve?
C'mon Anandtech, I know you can do this right - if you're going to benchmark and discuss the USB 3 capabilities of these boards, do that!
USB devices have always had some level of overhead compared to SATA devices. USB 1.1 devices never actually achieved 12Mb/s but were closer to around 8Mb/s (1MB/s) maximum. USB 2.0 is rated at 480Mbits, which translates into 57.22 MiB/s (remember that Windows uses 2^20 MiB for transfer rates instead of 10^6 MB). However, there's overhead and the maximum sustainable transfer rates on USB 2.0 tend to be closer to ~38MiB/s. At that speed, most conventional HDDs can easily saturate USB 2.0, but they're not significantly faster (i.e the fastest HDD is going to be in the realm of 100 MiB/s).
Move to USB 3.0 and the maximum theoretical data rate is 572 MiB/s, but there's still overhead that limits performance to 4 Gb/s instead of 4.8Gb/s, and it looks like transfer rates of up to ~380MiB/s will be considered "ideal". Even the fastest SSDs aren't going to come close to 380 MiB/s right now, as they would need external enclosures and SSDs with SATA6G support.
With USB 3.0, a 1TB HDD transferring data at around 93MB/s is very close to the speed of a 1TB drive connected via SATA, so the HDD is now the bottleneck for USB 3.0. If Raja tested with a fast SSD, we could see if USB 3.0 can get up to ~250 MB/s, but we can't test if it can achieve any better than ~43% of the theoretical throughput without SATA 6G.
I'm sure we can come up with a way to challenge USB 3 and make sure the boards are actually performing to the spec. There are two USB 3 ports on the motherboard, how about writing large files to SSDs attached to each one? How about copying a file from one to the other? How about streaming off of a USB 3 camera or two?
1) Couple of ITX reviews are in the pipeline - including the Zotac H55 and Intel M-ITX 'Jet Geyser'. The Intel review is what I'm working on right now (board is due for release mid-march with an MSRP circa $125)
2) We've currently got 3 Intel boards in lab due to be reviewed. The Intel 'Jet Geyser' is first. The two other boards in-lab from Intel are the P55 M-ATX and ATX models, reviews for both should be up sometime in March..
We're focusing on the Intel and Zotac M-ITX boards at present,so it will be a few weeks before we can look at the models you mention (will def try and add them in if possible somewhere down the line).
Based on past experience, how do you expect the Intel boards to compare to the other manufacturers?
Strange that Intel is only using the H57 in the Mini-ITX board and not in the larger two. Seems counterintuitive, like in many cases aren't you more likely to want the RAID features in a larger case that can hold more drives, therefore accommodating a larger motherboard also? Maybe that's just my preference or assumption.
I always wondered how the referance board's preformed. And now as I get older, would consider one if its more stable out of the box then some 3rd party boards.
When I was younger, ie worked less, had more time to tweek systems, an issue or two, or even a few wasnt a big deal. Now I'd like to be installing the OS as soon as everything is screwed togther, and not have to worry about some setting not working.
Thank you for finally putting these mobo companies in their place. The inability to release these boards with even rudimentary function has gone on long enough and by you not only mentioning the issues, but more importantly, by not giving awards is more than a small step in the right direction.
While I did not fully read each and every page (I'm not currently in the market so the overview of the platform and conclusions were relevant), your analysis seemed very thorough.
It still doesn't seem like any of these boards is ready to build a ROCK solid stable system around like in the era of the old BX-chipset days. That said why did you not compare the intel H55 board. I would like to know how good and stable they are. I don't just mean that it can run a CPU benchmark, but how does the network and sound and usb handle wake from suspend, memory compatibility etc. I am leaning towards the intel H55 board.
why is it so hard to choose a good board for this platform, that it takes two articles with almost 10 pages each in a graduated site like this to aid people in the choice???
i say this platform is bogus, by design. not that the system is faulty and etc, but everything was made "so fast" that intel did not test the design properly... so, i repeat what i said in the first article: I, as consumer, will simply "skip" this generation of CPU's. maybe in the next iteration everything will be more polished.
There's probably some blame on both sides (Intel & sub-vendors), for some of the required fixes. Stuff like AHCI, and testing if four memory modules boot or are stable are a responsibility of the sub-vendor prior to public BIOS/board release though.
It's always wise to wait a few months before jumping on a new platform. There is room for improvement in some areas as pointed out above, but the advice to wait is ALWAYS sound.
The reason it took 10 pages per article and the time it did is mainly down to how much testing is involved (stability, peripherals, benchmarks + the time it takes to pen it all down).
Any idea how we can get OEM vendors to update their bios's for longer periods of time? Why is it we have to pay a fee to Phoenix for updated bios' when the OEM is no longer willing to carry the updated version?
They keep doing the same thing over and over again. If they have that many problems with their most popular products imagine what happens to their lesser known parts.
We've finally noticed a trend in a batch of PCs at work that keep crashing or failing to boot with Asus P5LD2-TVM-SE-SI-1394 mobos. They've admitted it finally but we've had this issue for 2 years already. We're lucky our OEM supplier can put big pressure on them but if we were a small company then we'd stand no chance of even getting an acknowledgment
I don't know if you are still monitoring this article's comments, but I was wondering if you have any idea if the latest Gigabyte H57M-USB3 F5 BIOS fixes the AHCI problems you were reporting? I am looking for a mATX motherboard with USB 3.0 and this motherboard with it's PCIe x4 slot that could be useful for a future USB 3.0/6Gps SATA expansion card seems like a good option if it's niggling issues have been worked out.
The ECS H55H-V1.0, convince me. It has the best price and everything I need. A PC for my wife, is economic and profitable MB. It is not behind the other more expensive solutions. I just have to choose the correct CPU.
Do you know the difference between the two fabrication revisions? As a solution provider, I'm in direct contact with Gigabyte sales in my region. They tell me: "We have an unique power design of USB port to charge iphone/ipad from S1 to S5 mode. Rev. 2.0 one improves the convenience. But I ensure there is no difference of functions. Please consider this a running change."
however, as a hardware Engineer, I see 1.0 -> 2.0 as a major revision increment, as opposed of 1.0 ->1.1.
If you look at the pictures of the GA-H57-USB3 rev1 and rev2 you can see a different IC (upper left of picture) installed on the boards between the revisions.
I don't have the board in hand yet and the sales tells me he will supply a rev1 board. This is why I'm asking any thoughts? if you still have the GA-H57M-USB3 on hand what revision do you have and can you tell me the chip part number (just under the "Designed in Taiwan" on the upper left of the picture)
Samsung, a respected brand keeps introducing mobile phones that match today’s speed and technology. Samsung has gone great lengths this time as it has received impressive reviews for its new Galaxy S 4G which happens to be the first 4G Phone with 21+ Mbps speeds.
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
56 Comments
Back to Article
ReaM - Thursday, March 25, 2010 - link
Hey, you promised to test Quadcores 860 and 750 on the H55. I am wondering how those run and overclock on that chipset.You promised in Part1!
It all leads to: P55 for quads or H55?
ReaM - Friday, March 26, 2010 - link
Never mind, I bought the p55m ud2 from Gigabyte, also thanks to you review for that board :)Thanks for posting this free tests
<3 <3 <3 <3 <3 <3 <3 <3 <3 <3
wysiwygbill - Monday, March 22, 2010 - link
Contrary to the original announcements the H55-itx does NOT support dual link DVI and the maximum DVI resolution is actually 1920x1200.RodEvan - Tuesday, February 23, 2010 - link
Thanks Raja for the excellent review - I've held back on purchasing a motherboard awaiting 'Part 2'. I'm looking for a relatively low power consumption board for a windows home server setup.I was actually pretty amazed by the Gigabyte idle power results - as equal to the the MSIboard (44 watt).
For much of the time the CPU will be used minimaly - only rising for the occasional media decoding task - so for me the Gigabyte boatrd seemed ideal - partcularly since for some HD media sata3 speeds might come in useful.
But the section on power consumption concludes
"Best overall power consumption figures belong to MSI, while boards laden with USB 3.0 features and SATA 6G trail in both idle and load situations."
The Gigabyte board (H55-USB3), quite remarkably doesn't appear to trail at all on idle.... am I missing something here?.
jed22281 - Wednesday, February 24, 2010 - link
what did you mean by:"particularly since for some HD media sata3 speeds might come in useful."
The board doesn't have SATA3 does it? Only USB3.
And what HD media would benefit from SATA3 speeds?
There isn't any HD media I'm aware of that'd come close to saturating sata2.
Or am I misunderstanding what your were trying to say?
Cheers
RodEvan - Wednesday, February 24, 2010 - link
No misunderstanding - I probably need to say a little more.I currently one one Sagetv media extender off the server. But am about to install two more. The system should ideally allow for 3 HD videos to be streamed off the same HD at the same time.
I haven't done the calaculation - but having the option for SATA3 seemed like a good idea.
I've also re-read the review a note that this board has "Gigabyte SATA 2 chip: 1 x IDE, 2 x SATA 3Gb/s (RAID 0, 1 and JBOD)"
and NOT 2 *SATA3 g/bs ... my misreading,sorry for the confusion.
Rod
michal1980 - Wednesday, February 24, 2010 - link
what are you streaming in hd? Blu-ray specs call for ~54Mbps Peak.3 times that is 162. Which is just barely over Sata1 spec, well under sata2. Sata3 @ 600Mps would allow for ~11 streams (not counting for overhead).
At that point I'd be more worried about drive/raid performace then saturating the bus.
Rajinder Gill - Tuesday, February 23, 2010 - link
Hi,Thanks for the feedback. The comment was made considering all three scenarios (Idle/full load/video playback), the MSI board came out on top in all three, hence the comment 'best overall power consumption'.
Hope that clears it up..
regards
Raja
NickCardwell - Tuesday, February 23, 2010 - link
I would love to see an Anand review of the Intel DH57JG and Zotac H55ITX-A-E motherboards. I am looking to put one of them (probably the Intel) along with an i5-650 into a Lian Li PC-Q07 case. I love to build powerful systems for others but stick to small and simple for myself.Ogopogo - Monday, February 22, 2010 - link
How do the Intel boards compare to these other boards? My main main criteria is stability.I don't know if a third installment of this upcoming but it would be interesting to see a test with a displayport H55
jackylman - Monday, February 22, 2010 - link
First, thank you for the detailed information on fan control. I've made a few comments about that in the past and it's nice to see that category now seems to be part of a standard Atech mobo review.I'm not in the market for a Clarkdale platform, but if I was, this article would be very useful.
Shadowmaster625 - Monday, February 22, 2010 - link
A lot of people are talking about new motherboards being released before they are ready. And for good reason. Why bother? Why not go and buy a well known and well established motherboard like the GIGABYTE GA-G31M-ES2L for $45 and drop in a E5200 and a quiet GT 220 or 5450? Overclock it to a modest 3 GHz and it will surely smoke this H55/H57 garbage in all the gaming benchmarks, for a LOT cheaper.I do not understand the value in this entire product line. Why do you not compare these with the option I just mentioned? I dont care about how intel wastes their monopoly advantage. If hardware from a year ago is cheaper and better than this crap they are shoveling out now, then it is your responsibility to tell us that.
TrackSmart - Tuesday, February 23, 2010 - link
Point taken, but as readers, isn't it great that we don't have to buy all the newest, most expensive hardware to find out how it runs? And the problems with it? We can just read articles like this one.The Anandtech folks *do* write articles showing budget parts that offer exceptional value (via overclocking or unlocking cores). This just isn't one of those articles.
Rajinder Gill - Monday, February 22, 2010 - link
Hi,I think the E5300 was benchmarked against Clarkdale here in our chipset/CPU launch articles.
http://www.anandtech.com/cpuchipsets/showdoc.aspx?...">http://www.anandtech.com/cpuchipsets/showdoc.aspx?...
regards
Raja
Taft12 - Monday, February 22, 2010 - link
Interestingly, a tasty OC on an E5300 will push the benches up towards the E8600 in those charts, that is to say, faster than Clarkdale.OK, so an E5300 won't get quite THAT far, but it shows you that Clarkdale is marginally better than Wolfdale at best and not at all worth the price.
lukeevanssi - Saturday, June 26, 2010 - link
if anybody want to know more about it so plz visit this link:-http://www.healthproductreviewers.com/force-factor...
there is a lot off knowledge about this product
TrackSmart - Monday, February 22, 2010 - link
Thank you for pushing Gigabyte on the AHCI issue! Can you ask them about their 790-series boards, too? I'm frustrated with the lack of AHCI support on my new GA-790-XTA-UD4 motherboard. There's a 30-45 second delay in initializing SATA hard disks when returning from sleep mode. This causes Windows 7 to blue screen. The only fix is to revert to IDE mode for all drives. I wasted 2 days trouble-shooting this only to find out its an unaddressed problem with AHCI support on this motherboard (and many others).Lukas - Monday, February 22, 2010 - link
This may not be the solution you're looking for, but it fixes the bluescreen at least:http://support.microsoft.com/kb/977178/#top">http://support.microsoft.com/kb/977178/#top
TrackSmart - Monday, February 22, 2010 - link
Thank you for the tip. I tried the hotfix.It's funny because the hotfix definitely prevents total operating system failure (i.e. BSOD). However, Windows takes up to a full minute to become responsive when resuming from sleep mode. Presumably the OS is waiting as long as it needs to for the SATA boot drive to become responsive again.
I will continue running in Native IDE mode for now, since losing 1% system performance is less irritating than waiting forever for my system to become responsive.
*** It would be nice if AMD or Gigabyte addressed the true problem, but I won't hold my breath. I still haven't heard back from their customer support about this problem and it's been a few days. Not even a "we'll get back to you soon" message. Nada.
Taft12 - Monday, February 22, 2010 - link
Great article Raja, I also appreciate the detail of the good and bad!I have a question for you or anyone else who might know - you mentioned ASUS dropping the MSRP of their H55 board at the start of the article... Where can I look up what the different vendors MSRP's are? Intel and AMD have made it quite easy to find out the 1K unit price of their CPUs on their own sites, but I haven't seen anything similar for motherboard vendors. Is there an authoritative, frequently-updated source?
7Enigma - Monday, February 22, 2010 - link
Taft,Why does the MSRP really matter? Unless you are looking to buy in bulk, I don't really see what the benefit would be. For instance the Intel 80gig SSD has been inflated for months now. It's still a great product, and I don't know why someone wouldn't buy something if the price was higher than MSRP. Many electronic products sell below MSRP, and IMO that does't make them any better or worse.
More importantly than MSRP in my opinion is shopping around to get the best price at any given point. 15-30min of googling around to various E-tailers will give you a good ballpark of the price.
Rajinder Gill - Monday, February 22, 2010 - link
Hi,Unfortunately, I haven't seen or heard of any such resource from any of the sub-vendors - I usually have to fire off emails to ask. I agree with you though, it would be nice to see the figures posted on vendor websites.
regards
Raja
willsonjone - Tuesday, March 22, 2011 - link
Samsung, a respected brand keeps introducing mobile phones that match today’s speed and technology. Samsung has gone great lengths this time as it has received impressive reviews for its new Galaxy S 4G which happens to be the first 4G Phone with 21+ Mbps speeds.Samsung Galaxy S 4G
Kid98 - Monday, February 22, 2010 - link
(ASUS is using a 188 amp capable VRM vs a guessed 110~120amp VRM on the Gigabyte H55/H57 boards, so we'd pick the ASUS boards for Lynnfield CPU overclocking)Looks like the units used to express the capabilities of the VRM's is incorrect, otherwise I am going to need a really large PSU to run them....
Kid
Swivelguy2 - Monday, February 22, 2010 - link
That's 100+ amps delivered at a Vcore of ~1 volt, not 100+ amps drawn from the 12v rail.Earballs - Monday, February 22, 2010 - link
I don't understand these Farcry 2 benchmark results.How is this score against the i5 750 being calculated? Did you test all these boards with that processor then average it? :S
If the first result is the 661@stock and the second the 661@4Ghz, why is the comparison to the i5 750 number not the same in both charts? What changed for the 750? The graphics settings appear to be the same.
Rajinder Gill - Monday, February 22, 2010 - link
The 750 was benched on the EVGA P55 FTW board. Same drives, same drivers, same GPU, same patch version of FC2. Not sure why Lynn (4core) does worse in this bench.later
Raja
Earballs - Monday, February 22, 2010 - link
Okay thanks, but that doesn't explain this:If the first result is the 661@stock and the second the 661@4Ghz, why is the comparison to the i5 750 number not the same in both charts? What changed for the 750? The graphics settings appear to be the same.
Rajinder Gill - Monday, February 22, 2010 - link
Hi,The 750 was OC'd to 4GHz in the second chart, which is why the figure is 5FPS higher than the stock number of 68fps.
I can update the graph in about a day with the 750 running in one of the H55 boards if you like - got the Intel JG on the test rig at present and it won't do 4G's with the 750 I suspect..lol
regards
Raja
Earballs - Monday, February 22, 2010 - link
Ok. It might clear things up to have the 750 labeled as OC'd in the second chart. Thanks for the replies.Swivelguy2 - Monday, February 22, 2010 - link
Okay, USB 2 transfers a gigbyte in 36 seconds. This is consistent with the rated spec of 480 Mbit/s. Isn't USB 3 supposed to be 10x faster? What's causing the limitation? Is it the implementation by Asus and Gigabyte (like attaching the USB 3 to too narrow of a PCI-e lane)? Is it simply the write speed limitation of the USB HDD used? If the latter, why not write to some flash memory to see if the results improve?C'mon Anandtech, I know you can do this right - if you're going to benchmark and discuss the USB 3 capabilities of these boards, do that!
Rajinder Gill - Monday, February 22, 2010 - link
Good point. In-depth analysis of USB 3 & external HDD's etc will be offered up in a dedicated article.later
Raja
JarredWalton - Monday, February 22, 2010 - link
USB devices have always had some level of overhead compared to SATA devices. USB 1.1 devices never actually achieved 12Mb/s but were closer to around 8Mb/s (1MB/s) maximum. USB 2.0 is rated at 480Mbits, which translates into 57.22 MiB/s (remember that Windows uses 2^20 MiB for transfer rates instead of 10^6 MB). However, there's overhead and the maximum sustainable transfer rates on USB 2.0 tend to be closer to ~38MiB/s. At that speed, most conventional HDDs can easily saturate USB 2.0, but they're not significantly faster (i.e the fastest HDD is going to be in the realm of 100 MiB/s).Move to USB 3.0 and the maximum theoretical data rate is 572 MiB/s, but there's still overhead that limits performance to 4 Gb/s instead of 4.8Gb/s, and it looks like transfer rates of up to ~380MiB/s will be considered "ideal". Even the fastest SSDs aren't going to come close to 380 MiB/s right now, as they would need external enclosures and SSDs with SATA6G support.
With USB 3.0, a 1TB HDD transferring data at around 93MB/s is very close to the speed of a 1TB drive connected via SATA, so the HDD is now the bottleneck for USB 3.0. If Raja tested with a fast SSD, we could see if USB 3.0 can get up to ~250 MB/s, but we can't test if it can achieve any better than ~43% of the theoretical throughput without SATA 6G.
Swivelguy2 - Wednesday, February 24, 2010 - link
I'm sure we can come up with a way to challenge USB 3 and make sure the boards are actually performing to the spec. There are two USB 3 ports on the motherboard, how about writing large files to SSDs attached to each one? How about copying a file from one to the other? How about streaming off of a USB 3 camera or two?michal1980 - Monday, February 22, 2010 - link
Firstly, thanks for ripping into these compaines.Secondly; Any plans on testing any of the itx boards coming out? Zotac?
Thridly. How come reference intel boards are never tested? It would be nice to see how intel mobo's are right out of the gate.
Rajinder Gill - Monday, February 22, 2010 - link
Hi,1) Couple of ITX reviews are in the pipeline - including the Zotac H55 and Intel M-ITX 'Jet Geyser'. The Intel review is what I'm working on right now (board is due for release mid-march with an MSRP circa $125)
2) We've currently got 3 Intel boards in lab due to be reviewed. The Intel 'Jet Geyser' is first. The two other boards in-lab from Intel are the P55 M-ATX and ATX models, reviews for both should be up sometime in March..
Hope this helps..
-Raja
The Wasp - Tuesday, February 23, 2010 - link
Raja,Any chance you'll be reviewing the new ECS H55H-I ITX motherboard?
How about the Giada MI-H55?
I'm trying to pick the best 1156 ITX motherboard, so it would be great to get some more in-depth info on all of the options.
Thanks!
Jon
Rajinder Gill - Tuesday, February 23, 2010 - link
Hi Jon,We're focusing on the Intel and Zotac M-ITX boards at present,so it will be a few weeks before we can look at the models you mention (will def try and add them in if possible somewhere down the line).
regards
Raja
The Wasp - Thursday, February 25, 2010 - link
Hello again,I don't know if it's available, but if you could include the Jetway NC97 that would be awesome too!
Jon
grazapin - Monday, February 22, 2010 - link
Based on past experience, how do you expect the Intel boards to compare to the other manufacturers?Strange that Intel is only using the H57 in the Mini-ITX board and not in the larger two. Seems counterintuitive, like in many cases aren't you more likely to want the RAID features in a larger case that can hold more drives, therefore accommodating a larger motherboard also? Maybe that's just my preference or assumption.
michal1980 - Monday, February 22, 2010 - link
Sounds good.I always wondered how the referance board's preformed. And now as I get older, would consider one if its more stable out of the box then some 3rd party boards.
When I was younger, ie worked less, had more time to tweek systems, an issue or two, or even a few wasnt a big deal. Now I'd like to be installing the OS as soon as everything is screwed togther, and not have to worry about some setting not working.
7Enigma - Monday, February 22, 2010 - link
Thank you for finally putting these mobo companies in their place. The inability to release these boards with even rudimentary function has gone on long enough and by you not only mentioning the issues, but more importantly, by not giving awards is more than a small step in the right direction.While I did not fully read each and every page (I'm not currently in the market so the overview of the platform and conclusions were relevant), your analysis seemed very thorough.
Congrats!
FATCamaro - Monday, February 22, 2010 - link
It still doesn't seem like any of these boards is ready to build a ROCK solid stable system around like in the era of the old BX-chipset days. That said why did you not compare the intel H55 board. I would like to know how good and stable they are. I don't just mean that it can run a CPU benchmark, but how does the network and sound and usb handle wake from suspend, memory compatibility etc. I am leaning towards the intel H55 board.marc1000 - Monday, February 22, 2010 - link
why is it so hard to choose a good board for this platform, that it takes two articles with almost 10 pages each in a graduated site like this to aid people in the choice???i say this platform is bogus, by design. not that the system is faulty and etc, but everything was made "so fast" that intel did not test the design properly... so, i repeat what i said in the first article: I, as consumer, will simply "skip" this generation of CPU's. maybe in the next iteration everything will be more polished.
Rajinder Gill - Monday, February 22, 2010 - link
Hi Marc,There's probably some blame on both sides (Intel & sub-vendors), for some of the required fixes. Stuff like AHCI, and testing if four memory modules boot or are stable are a responsibility of the sub-vendor prior to public BIOS/board release though.
It's always wise to wait a few months before jumping on a new platform. There is room for improvement in some areas as pointed out above, but the advice to wait is ALWAYS sound.
The reason it took 10 pages per article and the time it did is mainly down to how much testing is involved (stability, peripherals, benchmarks + the time it takes to pen it all down).
later
Raja
JonnyDough - Wednesday, February 24, 2010 - link
Any idea how we can get OEM vendors to update their bios's for longer periods of time? Why is it we have to pay a fee to Phoenix for updated bios' when the OEM is no longer willing to carry the updated version?semo - Monday, February 22, 2010 - link
They keep doing the same thing over and over again. If they have that many problems with their most popular products imagine what happens to their lesser known parts.We've finally noticed a trend in a batch of PCs at work that keep crashing or failing to boot with Asus P5LD2-TVM-SE-SI-1394 mobos. They've admitted it finally but we've had this issue for 2 years already. We're lucky our OEM supplier can put big pressure on them but if we were a small company then we'd stand no chance of even getting an acknowledgment
cactusdog - Monday, February 22, 2010 - link
"at least until SATA 6G peripherals are affordable"Western Digital have just released a 1T Sata 6Gbs drive and its not much more expensive than a Sata2 drive. A review on it would be nice.
Anyway nice roundup.
Ben90 - Monday, February 22, 2010 - link
Because it has the exact same performance as a SATAII driveRajinder Gill - Monday, February 22, 2010 - link
Thanks. You're right - I should have written SATA 6G SSD's..regards
Raja
ltcommanderdata - Tuesday, April 6, 2010 - link
I don't know if you are still monitoring this article's comments, but I was wondering if you have any idea if the latest Gigabyte H57M-USB3 F5 BIOS fixes the AHCI problems you were reporting? I am looking for a mATX motherboard with USB 3.0 and this motherboard with it's PCIe x4 slot that could be useful for a future USB 3.0/6Gps SATA expansion card seems like a good option if it's niggling issues have been worked out.RadMos - Wednesday, April 28, 2010 - link
The ECS H55H-V1.0, convince me. It has the best price and everything I need. A PC for my wife, is economic and profitable MB. It is not behind the other more expensive solutions. I just have to choose the correct CPU.hq5 - Friday, April 30, 2010 - link
How to modify the registry after installation to enable AHCI drivers on Gigabyte H57M-USB3 ?thx
taipeidman - Monday, June 14, 2010 - link
Hi Rajinder,Do you know the difference between the two fabrication revisions?
As a solution provider, I'm in direct contact with Gigabyte sales in my region.
They tell me:
"We have an unique power design of USB port to charge iphone/ipad from S1 to S5 mode. Rev. 2.0 one improves the convenience. But I ensure there is no difference of functions.
Please consider this a running change."
however, as a hardware Engineer, I see 1.0 -> 2.0 as a major revision increment, as opposed of 1.0 ->1.1.
If you look at the pictures of the GA-H57-USB3 rev1 and rev2 you can see a different IC (upper left of picture) installed on the boards between the revisions.
rev1 - http://www.gigabyte.com.tw/products/product-page.a...
- http://www.gigabyte.com.tw/fileupload/product/2/33...
rev2 - http://www.gigabyte.com.tw/products/product-page.a...
- http://www.gigabyte.com.tw/fileupload/product/2/34...
I'm asking here, if anyone knows...
I don't have the board in hand yet and the sales tells me he will supply a rev1 board.
This is why I'm asking
any thoughts?
if you still have the GA-H57M-USB3 on hand what revision do you have and can you tell me the
chip part number (just under the "Designed in Taiwan" on the upper left of the picture)
Taipeidman...
willsonjone - Tuesday, March 22, 2011 - link
Samsung, a respected brand keeps introducing mobile phones that match today’s speed and technology. Samsung has gone great lengths this time as it has received impressive reviews for its new Galaxy S 4G which happens to be the first 4G Phone with 21+ Mbps speeds.Samsung Galaxy S 4G
willsonjone - Tuesday, March 22, 2011 - link
HELLO