Gigabyte Z77X-UP7 Intel LGA 1155

[cadaveca]

Built with one of the most crazy power designs ever, featuring no less than 40 phases for CPU and memory, each capable of 60 Amps, the Gigabyte Z77X-UP7 is Gigabyte's new flagship model. This board does it all and has Quad SLI & CrossFire support. But do Gigabyte's words "A New Legend is Born?" ring true?

Show full review

 

[dj-electric]

Thx for the review fwend

IMO 399$ won't justify this board. A 300 price-point is what i expected. All in all, nice piece of FAPable hardware.

 

[Sasqui]

Gigabyte does seem to be gobbling up market share! That's an expensive board... I'd still choose the UD5H at half the price... unless I had 3 or 4 GPUs to run together of course.

Interesting about the USB 3.0 ports

UD3H has USB 3.0 only
UD5H has USB 3.0 and USB 2.0
OP7 has USB 3.0 only

 

[MxPhenom 216]

Do want!

 

[cadaveca]

Thx for the review fwend

IMO 399$ won't justify this board. A 300 price-point is what i expected. All in all, nice piece of FAPable hardware.

I did think about that. With all the add-ins in the box, plus the added features other boards neglect, like mSATA, and the eSATA and USB 3.0 panels, plus the PLX PEX8747, $400 seems pretty accurate.

It is also a 10-layer PCB, one the the thickest I've had my hands on yet for Z77.


It's also the most expensive VRM design on the market today. Those IR PowIRstages don't come cheap, and there is no less than 40 of them on this board.

Gigabyte does seem to be gobbling up market share! That's an expensive board... I'd still choose the UD5H at half the price... unless I had 3 or 4 GPUs to run together of course.

Interesting about the USB 3.0 ports

UD3H has USB 3.0 only
UD5H has USB 3.0 and USB 2.0
OP7 has USB 3.0 only

If you wanna run two GPUs, even, this board will be faster, since the PLX PEX8747 does seem to improve multi-GPU rendering in all instances. AT least, that's what the scores show. I have been testing boards with dual VGAs just beucase of the PLX PEX8747, actually. Boards without it are slower unless running a single GPU, and that's that.


Other PLX PEX8747-carrying boards don't all have the dedicated single-GPU slot, either, that negates the latency that the PEX8747 introduces with a single VGA.

Add in one of the best audio testing results ever, $400 seems very very fair.

 

[PatoRodrigues]

Gigabyte has a MultiCore Enhancement that really seems to give a nice boost....

I thought you would give it a 10 for this board. The price certainly didn't help. haha

 

[PatoRodrigues]

With MCE, the UD5H for example gets fairly close to ASUS Rampage boards using a i7-3960X with the 3770K.

Z77 + Hyperthreading = FTW

 

[cadaveca]

It's called "Legacy Benchmark Enhancement", and is not present in all BIOSes, but yes, it's here. I did mention this in the BIOS overview, too.

 

[drdeathx]

Dave, this board easily gets 5Ghz with a 3770K(temps high). Did you test it on air or water. I do agree with the PLX chip does not improve GPU performance in all instances as when i ran it against the MPower.

 

[cadaveca]

Dave, this board easily gets 5Ghz with a 3770K(temps high).

your point? if I had different clocks in every review, clock efficiency compares would not be possible.

If you wanna donate chips for me to burn up testing max clocks, I'll do it, but I'll need a chip for every board. The cost of such makes it not feasible, unfortunately. I'd love access to trays of chips...

The MPower did not match the Z77X-up7 at all, in my testing. The Z77X-UP7, using slower memory, was faster per clock, so there's no way the MPower competes.

I've yet to have a board that actual;ly makes a difference for CPU clocking...voltages might change, but for normal users, choosing one board over the other for overclocking the CPU isn't going to matter much.

When it comes to memory, or running sub-zero, then sure, board choice can matter.

Currently, boards are separated by the features offered, the BIOS, and slot layout.

 

[drdeathx]

your point? if I had different clocks in every review, clock efficiency compares would not be possible.

If you wanna donate chips for me to burn up testing max clocks, I'll do it, but I'll need a chip for every board. The cost of such makes it not feasible, unfortunately. I'd love access to trays of chips...

The MPower did not match the Z77X-up7 at all, in my testing. The Z77X-UP7, using slower memory, was faster per clock, so there's no way the MPower competes.

I've yet to have a board that actual;ly makes a difference for CPU clocking...votlages might change, but for normal users, choosing one board over the other for overclocking the CPU isn't going to matter much.


When it comes to memory, or running sub-zero ,then sure, board choice can matter.

The voltage was not high and I did not jam the temps at 5.1GHz. 4.6Ghz is meh IMO.

 

[cadaveca]

The voltage was not high and I did not jam the temps at 5.1GHz.

I do not have your chip, and not all chips behave the same = what you got isn't relevant, really. I don't mean to be curt about this, but I very purposely use "safe" voltage limits when overclocking to prevent degradation that may affect results with other products. IF I hurt the chip with this board, then the results with the next board are useless, IMHO, if I use a different CPU.

4.6 GHz is the max "Auto" clocks all boards offer for 3770K. To me, there is a reason OEMs came up with this number, so that clock actually has some basis in extensive testing done by the OEMs. Even MSI uses this exact clock for their "burn-in" testing on the MPower.


You will see though, if you look at all my reviews, even @ 4.6 GHz, different boards needed different volts to be stable. The clock is more than adequate to show those differences.

 

[drdeathx]

Point taken. The board does have more horsepower though for those pushing the limits. Just saying.....

 

[cadaveca]

Point taken. The board does have more horsepower though for those pushing the limits. Just saying.....

Oh, of course, that's a given. By no means am I maximizing clocks...to me, that takes LN2, and I don't do LN2 reviews currently. I'm more interested in relating usable results for all users. Not all 3770K's can do 4.6 GHz with reasonable voltage and temperatures, unfortunately. I had one stuck to 4.3 Ghz recently.

Really, what you are showing isn't the board's potential.. that's the CPUs potential. If a board affects CPU clocks, then it's because of ram settings, not the CPU-side of things.

What you've shown is more CPU potential than board potential. I have been told recently, that on HWBOT, binning for LN2 currently(and remaining competitive) requires a CPU boot into windows @ 5.0 GHz, with 1.25 V, on air, so even what you have shown is rather mediocre.

 

[drdeathx]

Oh, of course, that's a given. By no means am I maximizing clocks...to me, that takes LN2, and I don't do LN2 reviews currently. I'm more interested in relating usable results for all users. Not all 3770K's can do 4.6 GHz with reasonable voltage and temperatures, unfortunately. I had one stuck to 4.3 Ghz recently.

Really, what you are showing isn't the board's potential.. that's the CPUs potential. If a board affects CPU clocks, then it's because of ram settings, not the CPU-side of things.

What you've shown is more CPU potential than board potential. I have been told recently, that on HWBOT, binning for LN2 currently(and remaining competitive) requires a CPU boot into windows @ 5.0 GHz, with 1.25 V, on air, so even what you have shown is rather mediocre.

Than I don't have a binned chip and it does show the boards potential. Your statement is an oxymoron then.

 

[cadaveca]

Than I don't have a binned chip and it does show the boards potential. Your statement is an oxymoron then.

NO, it's not, because as I've said, board plays little in CPU clocking. Even the $89 Biostar TZ77XE4 is just as capable in that regard.


Do keep in mind, that at this point, I've tested well over 20 Z77 Express boards. CPU clocking is only an issue if voltages and loadline settings are not possible in BIOS, bar none, in all boards I have tested.

Do keep in mind too, that I test with more than one chip. I only include results with on echip for review, but I do test with both 3770K and 3570K, since the stuff about memory clocking most people relate has not proven to be true in my testing. Yes, each chip has a different max memory multi it can boot, but voltages paly little role in that for me. I must be one of few users running 2666 Mhz ram, 4x4 GB, on 3570k and 3770k, with 0.95V VCCSA, and 1.05V VTT..in other words.. stock IMC voltages. My approach is different, clearly.

 

[HammerON]

Another nice review Dave

 

[EarthDog]

Great review dave! Its a sexy sexy board!

RE: The overclocking... boards mean sooooooo little these days... I had the Asrock Z77 Extreme 4 to 5GHz and its a $110 board. The difference in some boards however is how much ACTUAL voltage (MM not software) is needed for a given clockspeed. THAT tidbit is interesting to me at least as it shows the efficiency of the power delivery bits vs other boards.

Sorry if I missed this but...

What was the max bclk?
What was the highest mem multi that would boot?
Do you use the auto voltage for 4.6GHz that you clock to?

EDIT: Holy cow.. $399? How is that a deal? Id much rather have the MVE for $30 less at this time (more features, no ram issues to name a couple of things).

I notice in the conclusion you mention the addition of quad SLI/Crossfire..Cant a MVE do quad SLI/Crossfire too? Also, power consumption is "great" in the conclusion but was the highest out of all the Z77 boards you tested...Did I misread something?

 

[cadaveca]

EDIT: Holy cow.. $399? How is that a deal? Id much rather have the MVE for $30 less at this time (more features, no ram issues to name a couple of things). Cant a MVE do quad SLI/Crossfire too? Power consumption is "great" in the conclusion but was the highest out of all the Z77 boards you tested...

I have the MVE here. I'll refrain from commenting about that compare until the review is live.

Power consumption, it should be higher, since none of the other boards use PLX PEX8747, yet it remains within a few Watts of the others that do NOT have this extra power draw source.

At stock, power draw was NOT higher than all others, which should be a good indicator of why I posted what I did.

I don't do BCLK testing, due to time limitations. I have like near 40 other points of compare via graphs of results in my reviews. How many tests do you put in your reviews? 16-20? All you use really is AIDA and SuperPi...

 

[EarthDog]

Fair enough on the features comparison (but uhh, board has been out for a bit now, there are no secrets, LOL!... but it does do Quad GPU's IIRC.

Oh, I didnt see any stock power results posted...

 

[cadaveca]

Fair enough on the features comparison (but uhh, board has been out for a bit now, there are no secrets, LOL!... but it does do Quad GPU's IIRC.

Oh, I didnt see any stock power results posted...

Those are on the installation page, always.

And yes, both do quad GPUs. That's not what I was looking at, however, more about actual slot bandwidth at each slot ,and how different users will use those slots based on number of GPUs.

And I know the board is out, but my opinion is not.

 

[EarthDog]

Those are on the installation page, always.

And yes, both do quad GPUs. That's not what I was looking at, however, more about actual slot bandwidth at each slot ,and how different users will use those slots based on number of GPUs.

And I know the board is out, but my opinion is not.

I see now.. thanks!

I was confused by what was actually written, thanks for your clearing up what was meant by that.

As always, great review...

EDIT: Hey, whats with the dig on what we test? Handbrake is important... sorry we miss that.
Being serious we test 16 different things, 7 items x2 (stock + overclocked) then max clocks in SPi 1M/32M. We dont count AIDA bandwidth testing as 4 items though! LOL! Our demo is a bit different than here too remember. We would test 3D stuff, but for the most part, results are within the margin of error so I personally stay away from that stuff.

 

[drdeathx]

Great review dave! Its a sexy sexy board!

RE: The overclocking... boards mean sooooooo little these days... I had the Asrock Z77 Extreme 4 to 5GHz and its a $110 board. The difference in some boards however is how much ACTUAL voltage (MM not software) is needed for a given clockspeed. THAT tidbit is interesting to me at least as it shows the efficiency of the power delivery bits vs other boards.

Sorry if I missed this but...

What was the max bclk?
What was the highest mem multi that would boot?
Do you use the auto voltage for 4.6GHz that you clock to?

EDIT: Holy cow.. $399? How is that a deal? Id much rather have the MVE for $30 less at this time (more features, no ram issues to name a couple of things).

I notice in the conclusion you mention the addition of quad SLI/Crossfire..Cant a MVE do quad SLI/Crossfire too? Also, power consumption is "great" in the conclusion but was the highest out of all the Z77 boards you tested...Did I misread something?

That is kind of where I am going. it is nice to know top overclockcks and voltages required. Whats the sense of overclocking if you just throw 4.6Ghz at each board. makes the overclocking part of the review... rather worthless since you are using the same chip. After all, this is a tech and overclock forum?

In my past experiences, this is what manufacturers love to show in overclocking section even if it is a $110 board. That manufacturers loves that shizza and if the Gigabyte board clocks on less voltage, they love that shizza. If I am reading the review, it is senseless to buy the Gigabyte board if the MSI Z77-GD45 overclocks as well. So showing why with actual chip perameters is the only way to show which is better in BOTH respects. The viewers that are extremists love to see those details.

 

[HammerON]

I would agree with you in that it would be nice to see how far he could push a board/cpu; however unless someone wants to start donating cpu's for him to use I don't see it happening

 

[EarthDog]

Ive been on the same CPU for my few Z77 mobo reviews. One even hit 5.150v benchable (Spi 1M, but still!). Just have to use your head, and he has a good one so it shouldnt be an issue...I also think he is under water too?

But like I said, the demo here is a bit different (tech forum, not full of benchers pushing the limit but a daily driver overclock group - we are like that as well, but a large part of our active segment are from our benchmarking team, which I dont think TPU has, so I understand not pushing it at all)