GA-EP45-UD3P rev1.6 for 3TB drives

[gto4ben]

<snip>
The attached bios contains the following mods:
</snip>

The .rar file would''t open with 7-zip (v9.2), is it perhaps corrupted? --Thanks

I tried 7-zip 2016-10-04 (64 bit) as well as WinRar 5.5 (64 bit) . Both opened the file without any problem. Can you try downloading again?

[Skeeter3]

<snip>
The attached bios contains the following mods:
</snip>

The .rar file would''t open with 7-zip (v9.2), is it perhaps corrupted? --Thanks

I tried 7-zip 2016-10-04 (64 bit) as well as WinRar 5.5 (64 bit) . Both opened the file without any problem. Can you try downloading again?

Thanks again, the fix was updating 7-zip to 16.04 (2016-10-04).

[Skeeter3]

Thanks again, the fix was updating 7-zip to 16.04 (2016-10-04).

The new BIOS booted fine with my E8400 processor and RAID1 SSD's.  Thanks gto4Ben!

[zamar26]

Hi gto4ben,

there are updates since the 09/13/2014 microcodes I previously used as well as updates available from Git-Hub.

Is it possible to use this BIOS for Rev.1.1 board? Did anyone try that? What's the worst case scenario, and how to get back from it? What are hardware differences btw rev 1.1 and 1.6 that can make the Mobo not recoverable with this BIOS?

If its risky, is it possible for you to update also this F11d BIOS for GA-EP45-UD3P rev1.0/1.1 Mobo with the same latest microcodes? I currently use your earlier mod of 2012 to that BIOS, and it runs great.

I also wonder if Intel already posted "Meltdown" and "Spectre" fixes for 775 socket processors?

[gto4ben]

I just checked and Intel still did not update the 775 processor microcodes with the "Meltdown" and "Spectre" fixes.

If you try to upgrade a rev 1.0 or rev 1.1 MB with a rev 1.6 bios, it will likely fail; however, these motherboards will recover because of the dual bios.

It's Easter today and I'm looking forward to a good meal and lots of chocolate with the family. I'll look into updating the rev 1.0 and 1.1 bios tomorrow. These revisions should use the same NCPUCODE as the rev 1.6 so the update should go smoothly.

[zamar26]

Thanks and have a great holiday!

Btw, since you still use these boards, what are your overclocking settings recommendations? I'm trying to overclock it with Q9400 and 8GB of Corsair Twin2X4096-8500C5 RAM. The problem is, this 1066Mhz RAM freezes Windows at BIOS defaults or any set voltages and values faster than 800Mhz, whether CPU is overclocked or not, despite no errors on Memtest. When I try to increase MCH core voltage, the freezes are getting earlier, so its near impossible beyond 1.2V, and it gets stable only at 1.05 - 1.1V range, but at 800Mhz firm. Any ideas? These RAM modules work great at 1066Mhz in ASUS boards, so just Gigabyte BIOS optimizations are missing. It looks like in-depth fine tuning of RAM Controller and RAM voltages shown by MemSet was seldom addressed on the web.

[gto4ben]

Here is the updated bios for the for
GA-EP45-UD3P rev1.0 using the F11d BIOS linked by Zamar26 with the following mods:

ICH10R southbridge - v11.2.0.1527 with the TRIM mod
RTL8111C ethernet - rtegrom266.lom
SATA2 chip - j1.08.01
ICHAACHI 1.2 original from a EP45-USB3P
Updated NCPUCODE

I checked if the CPU microcodes are still the latest using GitHub's MC_Extractor v1.16.1 r62. They are the latest.


With regards to overclocking this MB with a Q9400,and 8GB of the Corsair Twin2X4096-8500C5 RAM, I've tried in the past to tweak the RAM voltages and I found that when a DDR2 has a spec to OC at 1066MHZ using a RAM voltage of 2.0-2.1V, the are trying with the higher voltage to decrease the the setup times with faster rise times; however, this is self defeating because the higher voltage increases power dissipation, increases temperature, and slows them back down when hot, making the OC unstable. Better DDR2 RAM that can clock 1066MHz at 1.8V is required but the fast ones have significantly less storage and were really expensive. Aging also slows down the RAM so this also needs to be factored.

I  found that for my usage and stability needs, upgrading to the Xeon CPUs were a better option. They even  worked better than a stock setup of a Q9660 due to the lower thermal dissipation at rated speed. eBay currently has the x5482 around $30-$45. I kept my RAM at 800MHz and it runs stable without stressing any components. Of course you'll need to cut the CPU socket, buy the sticky tape etc. to make it work. The x5492 is more expensive but still reasonable.

If you're resigned to use the CPU and RAM you have, try adjusting the bios the DRAM timings (tRRD, tWTR etc), especially if you use only two sticks. The bios is conservative so you may be able to  reduce the number of clocks and get a mild performance increase. I filled up all my slots so the stock bios settings were set using the "Turbo" performance setting.

[gto4ben]

And for those with GA-EP45-UD3P rev1.1, 
here is the updated bios using the F10 version from Gigabytes website with the following mods:

ICH10R southbridge - v11.2.0.1527 with the TRIM mod
RTL8111C ethernet - rtegrom266.lom
SATA2 chip - j1.08.01
ICHAACHI 1.2 original from a EP45-USB3P
Updated NCPUCODE

[zamar26]

Here is the updated bios for the for
GA-EP45-UD3P rev1.0 using the F11d BIOS

Thanks for the mod and your suggestions. Will test shortly. Your mods should be reposted here for easier Google finding. I double checked, and I'm currently using this 2012 F11d BIOS referenced by the linked thread. Apparently, its a later 2012 release compare to my earlier 2011 release link, and its based on Gigabyte's own F11 beta BIOS, previously posted on their website and also included in the linked RAR, not on F10 BIOS they currently post. Would you be so kind to make your updates to that particular BIOS? Its marked as v1.0/1.1 on that forum. The notes say: Updated with... IRST 11.5.0.1347, Realtek 2.46, Jmicron 1.07.28 . Your updates have somewhat different ROM names, and of course dates.

I've 4 RAM slots filled, and so far managed to OC CPU by 30% with some effort, thus resulting in RAM running at 920Mhz instead of 800Mhz stable. It may not pass heavy stability benchmarks, but seems OK for daily use, and runs cool with a larger Cooler Master EVO212. Found no way to run it at 1066Mhz, and Corsair RAM guys only suggest 800Mhz and blame NB memory controller for that. I was wondering if Gigabyte managed to test this RAM and incorporated the best settings into BIOS like ASUS did for P5Q board were it runs very stable at 1066Mhz, but no, this RAM was not even listed as supported, and in fact PC couldn't boot into BIOS with Auto settings and 4 sticks until I upped Vram to 2.1 and dropped FSB/DRAM to 2.4B.

Would you care to explain in more detail, how replacing the older codes with new ones you added can benefit users? I was thinking once OS starts, anything BIOS related is no longer relevant. Then I found its not the case, for example fans speed is typically controlled by BIOS, and likely many other things hardware related. Can you shade a light on usefulness of BIOS updates, in addition to hardware stability due to manufacturer gradually finding the best settings for various chipsets and incorporating them into BIOS?

Btw, what software can identify this board revision number?

[gto4ben]

Here is the updated bios for the for
GA-EP45-UD3P rev1.0 using the F11d BIOS

Thanks for the mod and your suggestions. Will test shortly. Your mods should be reposted here for easier Google finding. I double checked, and I'm currently using this 2012 F11d BIOS referenced by the linked thread. Apparently, its a later 2012 release compare to my earlier 2011 release link, and its based on Gigabyte's own F11 beta BIOS, previously posted on their website and also included in the linked RAR, not on F10 BIOS they currently post. Would you be so kind to make your updates to that particular BIOS? Its marked as v1.0/1.1 on that forum. The notes say: Updated with... IRST 11.5.0.1347, Realtek 2.46, Jmicron 1.07.28 . Your updates have somewhat different ROM names, and of course dates.

I've 4 RAM slots filled, and so far managed to OC CPU by 30% with some effort, thus resulting in RAM running at 920Mhz instead of 800Mhz stable. It may not pass heavy stability benchmarks, but seems OK for daily use, and runs cool with a larger Cooler Master EVO212. Found no way to run it at 1066Mhz, and Corsair RAM guys only suggest 800Mhz and blame NB memory controller for that. I was wondering if Gigabyte managed to test this RAM and incorporated the best settings into BIOS like ASUS did for P5Q board were it runs very stable at 1066Mhz, but no, this RAM was not even listed as supported, and in fact PC couldn't boot into BIOS with Auto settings and 4 sticks until I upped Vram to 2.1 and dropped FSB/DRAM to 2.4B.

Would you care to explain in more detail, how replacing the older codes with new ones you added can benefit users? I was thinking once OS starts, anything BIOS related is no longer relevant. Then I found its not the case, for example fans speed is typically controlled by BIOS, and likely many other things hardware related. Can you shade a light on usefulness of BIOS updates, in addition to hardware stability due to manufacturer gradually finding the best settings for various chipsets and incorporating them into BIOS?

Btw, what software can identify this board revision number?

In order to avoid any confusion, can you please attach the bios you'd like modded to this thread that you are certain will work. I'm wary of a file that identifies itself being compatible for two MB revisions. Both rev1.0 and rev1.1 have their own bios on Gigabyte's website and I have not confirmed they are interchangeable by inspection or test.

My ROM naming convention are per my discretion to help identify what they are. The bios can accommodate longer ROM filenames so I took advantage of that capability.

IntelAMD update their microcodes whenever they fix a processor bug that's discovered in the field or new CPU models are brought to market that fit in the same socket (mobile versions come to mind).

For minor bugs, they usually don't publish what a new microcode revision fixes so I can't say with certainty what the benefit is. I recall that for the Gigabyte 775 MBs with Dolby Home theater, they initially implemented the bios incorrectly and Dolby would not work till the next bios revision. In this case, the benefit is evident.

For new CPU or DRAM models, the new microcodes allow them to be identified by the MB and used correctly (pin programming, Vcore, etc).

The MB manufacturer would need to insert the MB revision number into the bios in order to be read by software such as AIDA64 , CPUID  and others. I'm not sure why Gigabyte didn't do that since they already provide the MB model number.