Brolloks EVGA P55 LE review

[Deanzo]

LOL at the quote , I was in fact saying, @ 250 BCLK's, I could also boot with the mem at 2500Mhz cas9. Not that being able to boot is stable.

We all know thats what 1M is for.

 

[Shiggity]

There is a variable that has been overlooked in this thread.

Remember that the i5's have the PCI-e lanes inside the CPU. The frequency clocks of the CPU affect the PCI-e clocks and I believe that is what is limiting your overclock to 4.4GHz. The ratio between the BLK and PCI-e clocks get all messed up at really high cpu speeds.

This could ultimately be a limiting factor for these CPU's because I don't know if overclocking the PCI-e frequency really high is smart or safe. Don't want to blow up your new 5xxx's when you get em

I honestly do not think it's the board.



Looking ahead, this could also be a problem for SoC type designs and that new CPU with an integrated GPU on die.

 

[hokiealumnus]

The upper end EVGA boards (FTW, FTW 200 & Classified 200) have this feature:

Dual Clock Generators
Ensure the best signals
Dual CPU and PCIe clock generators mean you can push this board farther than any other

Time will tell if those clock any better because of it. It does seem to address your concern though. I did see someone at the eVGA forums from Korea post a screenshot with a 230Bclk on a FTW 200 IIRC. So maybe there is hope with that feature added.

 

[Alien432]

The ratio between the BLK and PCI-e clocks get all messed up at really high cpu speeds.

Any idea what the factor of that ratio is?

 

[mstersmith]

There is a variable that has been overlooked in this thread.

Remember that the i5's have the PCI-e lanes inside the CPU. The frequency clocks of the CPU affect the PCI-e clocks and I believe that is what is limiting your overclock to 4.4GHz. The ratio between the BLK and PCI-e clocks get all messed up at really high cpu speeds.

This could ultimately be a limiting factor for these CPU's because I don't know if overclocking the PCI-e frequency really high is smart or safe. Don't want to blow up your new 5xxx's when you get em

I honestly do not think it's the board.



Looking ahead, this could also be a problem for SoC type designs and that new CPU with an integrated GPU on die.

I am new here but have this same set up. I have yet to start OC yet. In the BIOS you have the option to change the PCIE freq directly. If CPU clocks effect the PCIE freq. could you not lower it directly and then indirectly increase it from the CPU OC? Just an idea, if you all think it will work I will give it a shot later.

 

[Shiggity]

I don't have any hard evidence to backup my claims, but I've read a couple articles that talked about how the PCI-e frequency doesn't like high CPU clocks.

Just something to think about. Brolloks will most likely play around with the PCI-e frequency when he has a chance, he's a busy man

I have no idea what the optimal clock : clock ratios could be, will probably take a bit of trial and error. But upclocking PCI-e frequency is very dangerous from what I've read, be very careful if you do it.

 

[Brolloks]

I did play with the PCIe speeds from 100 all the way to 110 and it only made the OC stability worse. Now that the PCIe lanes reside on the CPU it more so depends on how good your CPU is to yield a good OC, unlike the 775 platform where all you played with was the FSB and multi.

 

[Moltenlava]

You may want to check out some info posted in an anandtech review about how the PCI-E controller can affect OC's. As for their being a seperate option for PCI-E frequency EVGA's BIOS used to have an issue with showing what certain frequencies should be when you overclock the BClock, for example, if you set the BClock to 200 the x10 Memory Multiplier would still show a Memory Frequency of 1333 when the actual Memory speed will be 2000 due to the OCed BClock. They fixed that in later BIOS's for their X58 boards so this may just be a teething issue with the P55 showing PCI-E Controller speeds of 100 even though the BClock is OCed. It could be that frequency is separate and not related to the internal speed of the PCI-E Controller.

link:
http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=3634&p=19

You may have already seen this Brollock on the EVGA forums but thought i would post a link anyway just incase

link to post by Shamino about increasing PCH voltage with VTT.
http://www.evga.com/forums/tm.asp?m=100901208

That may be of no use to you but you never know.

 

[Brolloks]

Thanks for that Moltenlava, I did see Shamino's comments, I'm particularly cautious about the delta voltages across PCH and VTT and have noticed that PCH voltage increase as you increase VTT, so much so that it severs the data transfer from the SATA when you go above 1.5 VTT.

Raja1 over at EVGA forums has shown interest and trying to root cause some of my issues with this board. I will run the 750 that has proven to be a sweet chip on the Asus board as well on the EVGA today or this weekend.

 

[Shiggity]

Interesting, subscribed.

 

[Dapman02]

Quick question brollocks, is that a brown PCB or is it black.

 

[Brolloks]

Black PCB

 

[Dapman02]

So, do you recommend the ASUS over the EVGA

 

[Brolloks]

So, do you recommend the ASUS over the EVGA

In the mid range category thus far...YES !

 

[Brolloks]

Just to wrap up this thread:

The board is quite capable of yielding 4 Ghz on your 750, it did not work well for my 860, so don't be discouraged by some of my comments, I later found out that most 860's do not do well above 4.4 Ghz with HT enabled on air/water.

To be honest, I expected it to do the same as the X58 LE and the 920, the fact of the matter is that in general, my oppinion off course, the X58 platform is indeed overall stronger than the P55.

Hope this helps

Please PM me for any questions.