Hi.

If I may, when you guys are about to do a review on a product, can I suggest trying to do a little internet survey about potential problems BEFORE publishing your conclusions?

If you would have done that for the Extreme4 GEN3, you would have found very quickly a lot of forum posts about 2 major issues. I would like Asrock and/or you guys to comment about them:

1- Etron EJ168A USB 3.0 controller

Despite the fact this controller recently received the SuperSpeed USB Certification from the USB-IF, it is still plagued by problems such as random disconnect from external USB 3.0 hard drive. It also has issue with some web cam and other peripherals. Your USB device might work, or not. How would you like Russian roulette style USB connectivity with your brand new 185.00$ purchase?

Neither the latest 0.105 drivers or BIOS 1.10 from Asrock solved those problems. This is unacceptable for a brand new product. To Asrock defense, Gigabyte board with the same Etron controller have similar issues.


2- Integrated Intel video

Get this: If you use only the onboard video without a discrete card, the system reboot and/or crash randomly many time every day. According to Asrock tech support on those forums, the only current fix is to up de CPU voltage to 1.2V!!! This is an Asrock specific problem as I couldn’t find forum complaining about this from other vendors. Just Imagine someone not knowing about this major flaw and planning to use the Extem4 GEN3 with the onboard video only... Let the tech support / endless e-mail / multiple RMA begin…


Ramon

Dear Ramon,

When reviewing such products, we follow a pattern so we can compare performance with the other products previously tested.

Input like yours is appreciated. We will also communicate these issues to ASRock and let you know the results.

I have used the Etron USB 3.0 controller pretty much with USB 3.0 Flash Drives and rarely HDDs; the storage media used did not disconnect during that time.

Dear Ramon,

ASRock have replied for your first question and they said in case you have an Etron controller, you should update the drivers to the latest version on their website, 0.104 WHQL, not 0.105:

http://www.asrock.com/mb/beta.asp?Mo...xtreme4%20Gen3

People that use motherboards with J-MICRON USB 3.0 controllers should also update the controller firmware.

The second problem you have brought into question is still under investigation.

Hi.

Thanks for taking time to consider my request.

I didn't set the background as of why I made this post. Plain simple: This board (Extreme4 GEN3) is by far the best value for a 1155 CPU. At 185.00$, nothing comes close to it. I was about to buy 4 of them for me, my mother, my brother and one friend. Why build 1 PC when you can build 4? LOL. Then I started looking for issues and stumbled on a lot of forums post to finally look for reviews and found yours.

I'm glad you didn't seem to have problems with the Etron controller. BTW, Etron own website advertise version 0.105 of their WHQL driver: http : // **trip w dot** etron **dot** com/SystemICs.php . I guess Asrock is a little slow to update their website. Many have reported this driver (0.105) didn't solve their issues. Apparently, there are suspicions that the chip itself might overheat and only then exhibit problems. May I suggest you test again but with long, heavy and sustained transfer from fast internal hard drive to the fastest external USB 3.0 device you have? Forget about USB key, not enough bandwidth to stress the chip apparently. Also, please try to connect a bunch of different USB devices (USB 2.0 and 3.0) on the USB 3.0 port and see how they behave on the long run (say at least 2 weeks of daily work).

As for the onboard video, I really wish all this would only be a bad and unfounded rumour. Considering the impressive amount of post on different forums, I guess it must be real. Very sad because out of the 4 PC I wanted to build with this board, 3 of them are to be used with onboard video.

In conclusion, I think Asrock should very quickly come out with a revision 1.1 of this board with 100% fix for the Etron controller and onboard video issue. As it is, this board is broken and do not deserve to be sold as a fully functional product.

NOTE: I hope you didn’t get a “cherry picked”” board form Asrock that is somehow immune from those problems and not representative of the performance of a retail product. If you have doubt too, I strongly suggest you source a retail board and test it.


Regards,

Ramon

Hey Ramon,

I will let you know if I'll get any more information from ASRock.

Regarding the Etron controller, I do not think that I have got a cherry picked motherboard sample, since this part is not manufactured at all by ASRock, but only integrated by them on the motherboard.

Since you have asked again, I have transferred right now 65GB of data on a portable 2.5'' USB 3.0 HDD, and no errors were encountered (data was transferred with about 83MB/s).

What I think is that the problem manifests depending on what chipset you have in the enclosure, with some of them "liking" to communicate with the Etron controller, some not. More than sure, this incompatibility should be addressed with a newer driver version.

65GB of data at 83MB/s seem real good. And no disconnect from the external unit? This is good news. Lets hope my Western Digital My Book 3TB USB 3.0 will do as good as your test then (assuming I'll buy the Extreme4 GEN3).

As for the controller chip inside external USB devices having an impact on compatibility, it seems to be a logical point at first, but I have a counter argument: Why NEC, Renesas and pretty much all the other controllers out there have little or no issues at all compared to the Etron one? To support my point, many of those forum posts mention RMAing their Extreme4 in exchange for an Asus, MSI or Gigabyte with USB 3.0 controller from different vendors (every other devices and peripherals remained the same except the main board) and reported their USB devices now work properly when attached to the USB 3.0 port of the new board. That kind of single out the Etron controller as the culprit. The other alternative is Etron just had a bad batch of controller and that it's been fixed. Or maybe Etron drivers are indeed worse than half baked. I'm glad your board cleared with success a single 65GB transfer challenge. Still, I think a long term test with multiple USB devices is the only way to clear the air once and for all.

Cherry picking it quite easy for them to do. They order those Etron chip by the thousands. You simply bench them prior to integration on the motherboard and choose the one with the best characteristics: Lowest operating voltage, temperature and electromagnetic emission. Highest speed and switching capability under tighter than usual timing. If you do that, no mystery, the "selected" chip will perform way above the average of their retail counterpart. Especially if you can choose among thousands! And I'm not pointing fingers at Asrock here; everyone in the industry is doing it so they (Asrock) need to do it too to keep up. The only problem with this "strategy" is that if a major flaw appear and forums are filled with complains, this flaw will not show up in reviews like yours. Then people like me start asking questions about cherry picking and why your "sample" doesn't exhibit the same problems as many others have reported...


Can't wait to have your conclusions on the onboard video card.


Thanks,

Ramon

After doing some checkups on different websites, I have found that most freeze issues from the IGP were driver related, as seen on TweakTown:

http://forums.tweaktown.com/asrock/4...ceived-26.html

This guy also mentions that version 105 for the Etron controller is indeed causing some problems.

An updated driver pack can be now found on Xtremesystems Forums:

http://www.megaupload.com/?d=RPXVOLOT

However, as soon as I will get some information from ASRock, I will post in this thread.

In my tests, with Virtu Disabled, but connected to the motherboard DVI connector, which means that I was only running on IGP, but with the dedicated VGA still connected, I ran some benchmarks and I did not get any BSOD. From my experience with the H67, I was surprised to find out that while running on the IGP, I needed more voltage for the VDIMM in order not to get crashes in 3D benchmarks, even if the system was stable with Prime95 and memtest86 were not detecting any errors.

I'm glad you were able to confirm my findings.

Concerning integrated video, the posts I came across were really specific about using the on board video ALONE, with no discrete card installed in the system. That could be a BIOS issue where the IGPU is not feed enough voltage, specifically when no discrete card is detected. It could be that when using a discrete card in conjunction with the IGPU, even if you use the IGPU output, the BIOS setting might be OK under that specific circumstance. It could also be a driver issue. I recommend to emulate the same conditions to better the chances of reproducing the error. Try removing the discrete card, uninstall its drivers, reboot and run some 3DMark loop over night directly on the IGPU. In principle, you should wake up the next morning to a blue screen or a rebooted system.

I'm waiting till October 12 to make a move. Bulldozer comes out that day. We'll know if its value is better than a 2500K and it could mean price drop. If I end up choosing and Intel platform, I really hope that all doubts about the Extreme4 GEN3 will be dissipated by then.

Ramon

Dear Ramon,

By recently setting the Primary Graphics Adapter to IGP and letting the allocated memory for it inside the UEFI to Auto, I booted Windows 7 and uninstalled the Nvidia drivers (with the dedicated graphics card not installed), rebooted again for uninstalling the Virtu software too.
After making sure that I was using the latest version of the IGP software from Intel's website, I have went on and launched 3DMark 2005 and set a custom run of 20 tests for each type. The total run time was about 4 hours and a half and during this time the system did not hard reset by itself, lock or BSOD.

Hope you are satisfied with the result.

Stefan

I'm really glad you were able to run all the necessary tests to address those 2 concerns about the USB 3.0 controller and the on board video. It seem that at least some Extreme4 GEN3 currently available are not affected with the same issues reported around the web and work just fine, which is good news.

I wonder then if it could be related to a specific assembly line bug. I've heard some stories in the past where the performance/reliability of one board assembled in Taiwan was way better than the exact same board assembled in China. Quality control is not the same everywhere it seems! I would guess the sample you got actually came from Taiwan!

Anyway, thanks **A LOT** for taking the time to go to the bottom of this!

One last thing that may influence my final choice: SATA 3 performance. I came across the only other review of the Extreme4 GEN3 (From Bit-tech) and they've found SATA 3 performance for the Extreme4 GEN3 to be **WAY** below competition for both the Intel SATA 3 or the Marvell SE9120 SATA 3. I hope it's only a fluke or a bad test environment (bad driver, OS, SATA cables, etc...). I'm planning to buy a Crucial M4 128MB SSD drive as my boot drive and obviously would appreciate to get all the performance I can from such an expensive device.

The Marvell SE9120 is hooked to the system through a single PCI-e 2.0 lane, so I can understand the lower than expected performance: SATA 3 top out at 6Gb/s but a single PCI-e 2.0 lane only goes to 3.9Gb/s (500MB/s * 8 / 1024 = 3.9Gb/s) or only 65% of the expected bandwidth. The only way to choke the Marvel chip would be to hook 2 X SSD in RAID strip to rally get a bottleneck from the single PCI-e lane interface. It's either that or the Marvell chip is simply not performing at the same level as the Intel controller. Even with that limitations, not too many devices currently available on the market goes above 500MB/s anyway. On the other hand, it shouldn't happen at all when using the Intel SATA 3 ports which have full DMI bandwidth to the CPU.


Best regards,

Ramon

You're welcome!

Had to register just to post on the Asrock Z68 review, where Stefan did a great job of answering questions in the comments, and tested further to try and resolve/find issues...great job there!

cheers, from Brisbane, Australia

Many thanks for the appreciation!

I have found an interesting review about this EtronTech USB 3.0 Host Controller:

hardcoreware.net/gigabyte-ga-z68a-d3h-b3-motherboard-review/9/

Quote:
"This board is the first time I have seen the Etron EJ168A controller, and I have to say, it sucks. All throughout testing, I would have problems transferring files to and from the test system, and completing benchmarks. It would transfer fine for a while, but would randomly get “stuck” running at 1 MB/s or so. It wasn’t actually crashing, it was just lagging really bad. This happened across various USB 3.0 devices (although never came up when I tried a USB 2.0 device in the port). I have tried the latest drivers from the website, but to no avail.

So it could be that I have a bum board here, or it could be that the Etron controller just isn’t ready for real world use just yet. Things may improve with drivers in a while, but for now, it is not usable."

That's exactly in line with what many others have been reporting in many forums. Not only this USB controller is systematically 15-30% slower than NEC/Renesas implementation in benchmark, but it's also buggy and unstable as hell.

What bothers me the most is that the latest Etron drivers, V0.105, has been available for a long time (way over a month) with no update since. Asrock BIOS 1.1 for the Exteme4 GEN3 has also been available for many weeks now. Despite both drivers and BIOS improvements list mention better USB experience and supposedly solving some of the issues, the fact is we are still a long way to call this solution "reliable" and "competitive". Lets hope both Etron and Asrock up their game really quick to put this embarrassing situation behind them.

The only sign of hope is that recently, some board seems to be somehow unaffected, like the one used by MadShrimp for their review. Then again, that board was supplied by Asrock themselves, so credibility wise, its not a very convincing argument.

Ramon

This Etron USB 3.0 implementation works correctly with particular type and brand of USB 3.0 devices. It happenned that Madshrimp has tested the USB 3.0 ports with that kind of device. May I suggest to test with an external HDD enclosure with JMicron JMS539 chipset and you will see the incompatibilities.

Tests shall include several types and brand of USB 3.0 devices, not only one.

In my view reviewers have a great responsibility when they recommend a product without extensively testing it.

The funny thing is that until driver version 0.101, using Windows 7 64, there were random BSOD pointing to the Etron driver, even when the USB 3.0 ports were not used. See this forum: sevenforums.com/crashes-debugging/154631-random-bsod-driving-me-crazy-8.html
It is a shame.

Now that you mention it, I remember some complaints did in fact revolve around some specific external USB devices and JMicron was indeed involved.

And I agree on the part that hardware review site giving award to products should do thorough investigations.


Ramon