![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
Gigabyte Z87X-UD5H LGA 1150 Motherboard Review @ [H]
- Thread starter FrgMstr
- Start date
I dug up the firmware on my board going into the marvell raid settings.
Firmware 2.3.0.1050 <---that is way newer then 1.0.0.1033
Bios 20291.24369.82.65535
With luck before F8 is final maybe Gigabyte will have a fix for this as it seems isolated but still prevalent.
Firmware 2.3.0.1050 <---that is way newer then 1.0.0.1033
Bios 20291.24369.82.65535
With luck before F8 is final maybe Gigabyte will have a fix for this as it seems isolated but still prevalent.
Yeah, it's a USB3.0 issue. You can change bios settings to enable usb3.0 at boot or something similar (if you could get into the bios, haha).
A keyboard connected via USB2 will work tho, so if you have an old USB2 bracket, or USB2 ports on the front of your case you can connect those via one of the USB headers on the mobo and plug your keyboard in, this will let you get into the bios and change what you need.
You need to use a USB 2.0 port to enable legacy / USB keyboard support for USB 3.0. And I'll tell you that this doesn't always work. Some devices just don't play nice with it. Normally motherboards that use USB 3.0 HUBS for multiplexing the Intel ports like the ASM1072 are normally fine. Dedicated controller based ports like the ASM1042 and VIA controllers normally won't work in BIOS / DOS type environments. It's the one real downside to dedicated USB 3.0 controllers. (Besides cost.)
So this is a bios bug? The only USB 2.0 ports are on my front panel. Those should work?
What a hassle!
Yep, plug the KB into the front USB 2.0 ports!
Edit: And again, once you can get into the BIOS and there should be a setting for "Enable USB3.0 at boot" or something very similar, which will hopefully let you move your keyboard to a USB3.0 port on the back...As Dan D said, sometimes it still doesn't work.
Last edited:
Swell. The other issue I had was with memory stability. I had two identical pairs of 1600mhz cas 8 ram filling all four banks. Each pair bought separately. Know that's a no no sometimes. Was getting random bsods until I backed the ram speed down and disabled xmp. Then upgraded to f8 beta and manually set ram timings and stable again. Thing is I read setting XMP on with two separate kits even identical will cause instability. Anyone know about this?
Thanks.
Thanks.
It does. Some keyboards behave correctly with USB 3.0 ports and some don't.
What are your thoughts on the SB package? That is one reason I chose this board over the UP4H plus the extra NIC.
The onboard Realtek is weak, even at 100% volume it's more pathetic then I thought it would be.
Firing up and turning on the SBX pro studio does gain some volume and sound is more rich, still not ear bleeding loud though.
My 10+ year old turtle beach card in my other computer is too loud at 100%
Guess it boils down to a dedicated card is always better then onboard even if considered high end.
The onboard Realtek is weak, even at 100% volume it's more pathetic then I thought it would be.
Firing up and turning on the SBX pro studio does gain some volume and sound is more rich, still not ear bleeding loud though.
My 10+ year old turtle beach card in my other computer is too loud at 100%
Guess it boils down to a dedicated card is always better then onboard even if considered high end.
I don't compare a lot of actual sound cards as I really don't use them all that often. But dedicated hardware has the potential to be better than any onboard audio. Otherwise what's the point? In any case you'll get weak audio without an onboard amplifier and sufficient power to drive it. I've found any board that has this to be enough for me in terms of pure volume but some people aren't happy until their ears bleed. If your one of those folks then you'll probably always be looking for a dedicated sound card.
This board is supposed to have a built in amp if you use the front headphone header, they claim it's for more sound for gaming and headphones.
I see zero difference whether I have mine plugged into the front or rear jack.
Maybe my lower end Sennheisers just can't take advantage of this which is a possibility.
The SB package does help quite a bit though.
Another old build of mine with a Gigabyte UD3R this package was available to download but it was for 30 days then you had to purchase. Thankfully they give it to you for free [albeit you do pay more for this board so it's technically included].
Guess the debate is still would it have been better if Gigabyte went with ALC1150 which has just a touch more SNR or would it make any perceivable difference.
I see zero difference whether I have mine plugged into the front or rear jack.
Maybe my lower end Sennheisers just can't take advantage of this which is a possibility.
The SB package does help quite a bit though.
Another old build of mine with a Gigabyte UD3R this package was available to download but it was for 30 days then you had to purchase. Thankfully they give it to you for free [albeit you do pay more for this board so it's technically included].
Guess the debate is still would it have been better if Gigabyte went with ALC1150 which has just a touch more SNR or would it make any perceivable difference.
Last edited:
Intel recommends (see table 47)RAM of no higher than 1.5 +/- 5% but it seems 1.35v RAM modules are supported.
However, some people will be quick to point out that they've been running 1.65v RAM in their SB/IB/Haswell systems without any problems thus far.
Your call I guess - I personally have not seen anyone kill their CPU by running 1.65v RAM.
But on the safe side I avoided anything higher than 1.5 when buying my kit.
However, some people will be quick to point out that they've been running 1.65v RAM in their SB/IB/Haswell systems without any problems thus far.
Your call I guess - I personally have not seen anyone kill their CPU by running 1.65v RAM.
But on the safe side I avoided anything higher than 1.5 when buying my kit.
I've been running this same 4770K I've got on the bench with 1.65v modules at 1866MHz-2400MHz speeds without any issues since the CPU launched. I'd like to point out that I probably throw more voltage and idiotic settings at this thing than people would generally do if it was their money which purchased the items.
No problems yet.
Sorry, to clarify:
The 2nd row of the table also says that 1.35v RAM is also supported +/-5%.
I've been running this same 4770K I've got on the bench with 1.65v modules at 1866MHz-2400MHz speeds without any issues since the CPU launched. I'd like to point out that I probably throw more voltage and idiotic settings at this thing than people would generally do if it was their money which purchased the items.
No problems yet.
Well, that was quickly pointed out!
At least we know these chips are durable
-- or, at least more durable than Intel rates them for.
Intel CPUs are always durable. I've been overclocking since the 486 days and building systems since the early 1990's. I've built hundreds of machines and serviced thousands as a repair technician. In that time I've seen probably a half dozen bad Intel CPUs total. Most were socket 423 and 479. And the thing is, the bulk of the CPUs that died went when the motherboards fried as well. I've seen only one DOA Intel CPU and two that died on their own with common / non-overclocked use. I've never damaged one through overclocking.
GPUs on the other hand I've killed a lot of.
No.
GPUs on the other hand I've killed a lot of.
No.
The issue is that some boards just aren't well made and their actual voltages may range well outside the 1.65v setting. You have to get to know each board and watch the voltages. Some will show a setting of 1.65v but show actual readings as high as 1.67 or even 1.68. In which case you need to manually adjust downward to ensure the RAM runs within the right voltage spec. And even boards which show spot on readings may be lying to you. You have to put a volt-meter to them to be absolutely sure. Unfortunately most of the boards that are probably guilty of being built that badly don't have v-check points on them.
I wouldn't call 1.65v gambling at all. You just have to look carefully at what your board is actually doing and adjust it accordingly. Remember I do this all the time with different motherboards. MSI, ASUS, and GIGABYTE all have good voltage regulation from what I've seen. Fairly stable voltages, predictable vdroop and generally good component selection and VRM designs usually exceeding VRD specifications. The cheaper ASRock boards on the other hand are another matter. Horrific vdroop, terrible voltage stability and components which are often unmarked and unidentifiable origins. Not all ASRock boards are like that as the Extreme series is actually pretty good. But their budget offerings are about as cut rate as I've ever seen.
They aren't the only ones but they are the ones which I work with the most which fall into that category. Reference 680i SLI and later NVIDIA chipset based boards, PC Chips, QDI and other manufacturers are or were really bad as well but we almost never review anything that bad as they simply won't pass the review process. Most of those boards get chucked into the trash as a result and you won't ever read about that.
I went the safe route and picked up Mushkin redline 1866 @1.5v.
AIDA64 shows the memory is in the top 5 in most benches against other high end systems.
From most reviews I've read 1600 is actually fine, 1866 is the sweet spot and anything over that you will see gains but in such a small % compared to price difference it's not really worth it unless you simply have deep pockets.
AIDA64 shows the memory is in the top 5 in most benches against other high end systems.
From most reviews I've read 1600 is actually fine, 1866 is the sweet spot and anything over that you will see gains but in such a small % compared to price difference it's not really worth it unless you simply have deep pockets.
I also ended up going the safe route this morning ordering 16GB (2x8gb) kit of Vengeance Pro 1866 cas 9-10-9-27 1.5v ram. Hope that 2 modules at 16gb will afford better overclocking vs 4 modules on all 4 banks. Anyhow can't believe how expensive this stuff has gotten lately. There must be a ram shortage out there since the fire.
That's because their voltage design was terrible. Their BIOS' also forced voltages to run way higher than necessary to make them appear stable, but this lead to premature failures. This is why they'd do well in reviews and dye days to weeks later.
Sorry to go off topic a bit to reminisce about my old Abit IN9-32X. It was one of the last Abit boards they made. Looked real pretty with its copper heatsinks and all but had issues like needing a cpu fan to post, soundblaster card bios-post irq conflict no post issues, SB snap-pop-crackle issues, sata port bus noise hd failure issues, and windows bsod issues. I still have that system and it makes a nice paperweight. I worked out the bugs after a couple years and it overclocked well enough despite having fsb holes. Good ole days!
This board must not be selling that well as it's on sale now at newegg for $190 plus a $20 MIR, that is less then what I paid 4 months ago with a combo promo
Possible the C2 boards are out now as well.
Oh yeah, they are tossing in one stick of ripjaw 1866 4G so you only need to buy another stick to make a match.
Factor that all in and the board is actually $128, that's a steal.
Possible the C2 boards are out now as well.
Oh yeah, they are tossing in one stick of ripjaw 1866 4G so you only need to buy another stick to make a match.
Factor that all in and the board is actually $128, that's a steal.
Hi,
My 4670K is running at 3.8ghz all across 4 cores. I thought turbo mode allowed for max of two cores. I ran Prime95 at 100% loaded for all 4 cores and the clock ratio did not change. Steady at 3.8ghz. Is this normal?
Edit: I'm running stock settings! Auto selected in most cases in bios with Xmp profiles on!
My 4670K is running at 3.8ghz all across 4 cores. I thought turbo mode allowed for max of two cores. I ran Prime95 at 100% loaded for all 4 cores and the clock ratio did not change. Steady at 3.8ghz. Is this normal?
Edit: I'm running stock settings! Auto selected in most cases in bios with Xmp profiles on!
Last edited:
You can enable and disable both how many cores are active and if you want turbo on or off.
I fooled around this myself just for fun.
Do note that if you have EIST speedstep turned off they will remain at full turbo so check that setting.
Check turbo boost and ratio, they should be on auto but if not you might have to force them on.
Been awhile since I've had to look at bios and my settings so I could be wrong about EIST, try it out.
Turbo can run on all 4 cores but it's stepped, first two should be 3.8, 3rd 3.7 and 4th 3.6 or something to that effect.
I fooled around this myself just for fun.
Do note that if you have EIST speedstep turned off they will remain at full turbo so check that setting.
Check turbo boost and ratio, they should be on auto but if not you might have to force them on.
Been awhile since I've had to look at bios and my settings so I could be wrong about EIST, try it out.
Turbo can run on all 4 cores but it's stepped, first two should be 3.8, 3rd 3.7 and 4th 3.6 or something to that effect.
I swear I have every thing on auto with first two cores at 38, and 37, 36 for last two with eist speedstep on. Will have to check. Also prime95 load temps are max 63 degrees for each core as reported by realtemp. Do I have room to overclock?
You apparently are not using the stock sink judging by your sig and your temps. I have not used reatemp but I use coretemp and stock I can easily his 212F with prime or intel burn but in real life scenarios while playing games like Skyrim, Mass effect, etc my cores are around 127F max.
Should have some headroom, can't see why not.
Should have some headroom, can't see why not.
No I'm not using the stock sink. Using the H100i but installation was tricky as the mounting was loose at first. Seems like the Gigabyte board was a little thinner than corsair expected and had to use rubber washers to get her tight. Was wondering though if you could reply back in Celcius?
Thanks.
So I checked the bios settings and all the cpu configuration setting are set to 'AUTO' including Eist which makes sense because CPU-Z reports the Cpu Frequency all over the place until I run Prime95. Then each core or all of them hit 38x100 instead of two cores at 38, then 37, and then 36 like Turbo is suppose to be for that cpu according to the Intel spec sheet for the 4670K. So the Gigabyte bios must be forcing all cores to a 38 multiplier. A cheat perhaps for benchmarking. Perhaps its the F8 Beta bios. Anyone with the UD5H care to comment what your cores multipliers are at under prime?
Thanks
Thanks
So you manged to flash the bios to F8 beta from that french driver website without any ill effects [besides the full throttle]?
What exactly is the F8 beta supposed to fix?
What program do you use to see each cores multiplier? I only know what I have set in bios but can't verify it.
Using prime 64 as it stresses all 4 cores my top boost is capped to the 4th core, the slowest.
What exactly is the F8 beta supposed to fix?
What program do you use to see each cores multiplier? I only know what I have set in bios but can't verify it.
Using prime 64 as it stresses all 4 cores my top boost is capped to the 4th core, the slowest.
I flashed to the F8 Beta that was on Station-Drivers cause I was having random multiple BSODs like every 30 minutes or less after the other all of a sudden. Figured out it was setting XMP on two pairs of the same ram bought separately might have been causing it but got desperate so I flashed. Actually things got stable again though I upped my system agent voltage 0.01v just to be safe along with turning of XMP and setting stable ram timing profile in bios. However since upgrading to the new Corsair Vengeance Pros at 1.5v at 2x8gb at 1867mhz vs the 1600mhz was before I can't tell the difference. No BSODs though. Had one weird crash or lock up starting up Hitman Absolution for first time but 99% stable at the moment despite all cores being at full throttle. Noticed when priming the cpu vcore goes as high 1.26 volts. Wonder if the 3DMark enhancement option has anything to do with it? I personally think its the bios rigged to behave this way for higher benchmarks for reviewers..
Who knows..
Who knows..