Xeon L5639 Overclocking on X58

[Burpo]

When I actually look at the 95 watt Xeon's vs the 130 watt i7 both at 32nm I don't see better overclocking.

You never do, but I'm at 200 bclk stable 24X7.. Lots of other people here are as well.. difference between a server CPU with 2 QPI links that costs $1400+ and a $500 desktop CPU with 1 QPI link..

 

[Makaveli]

thanks for the link.
I used to have the EVGA SLI (E758) but it had issues and EVGA replaced it for me with the FTW3.
Which encouraged me to get a 6-cores Xeon from ebay.

I hope I can at least get it up to 4GHz.
I have a huge case (ATCS 840) with lot of extra fans, and the NH-D14 to cool down the CPU.

EDIT: this seems to be a good guide: http://www.techreaction.net/2010/09/07/3-step-overclocking-guide-bloomfield-and-gulftown/

4ghz is a pretty standard overclock on these chips and the NH-D14 is a beast you will have no problems.

 

[Makaveli]

You never do, but I'm at 200 bclk stable 24X7.. Lots of other people here are as well.. difference between a server CPU with 2 QPI links that costs $1400+ and a $500 desktop CPU with 1 QPI link..

And i'm on a i7 130watt 200 bclk stable 24x7 so yes I don't see it.

From everyone in this thread that has been posting I don't see any of you showing me numbers better than what I can do with my chip! 5Ghz suicide runs don't mean anything.

So yes burpo i've yet to see evidence of these xeon being superior overclockers when you are comparing at 32nm!

And what good is that 2nd qpi link going to do in a single socket board????

And why do you keep quoting 5-6 year old retail prices like that is suppose to mean something?

 

[Burpo]

Doesn't matter what you say..They're different.. Intel even had different product codes for the X5600 series.. Any time you want to post your best stable cinebench scores. I'll be happy to school you further.. They were expensive for a reason.

"two product codes are used, 80613 for the UP desktop/server models and 80614 for the Xeon 5600-series DP server models."

http://en.wikipedia.org/wiki/Gulftown

 

[Venoom]

does anyone here have experience with 8GB memory sticks on the X58 platform?

I am hoping to get 24GB RAM with 3 sticks, and potentially 48GB with 6 sticks.

X58 is officially limited to 24GB (when you run it with a core i7, which you are supposed to do) but Xeons officially handle 288GB (Intel Ark)

 

[Burpo]

does anyone here have experience with 8GB memory sticks on the X58 platform?

I am hoping to get 24GB RAM with 3 sticks, and potentially 48GB with 6 sticks.

X58 is officially limited to 24GB (when you run it with a core i7, which you are supposed to do) but Xeons officially handle 288GB (Intel Ark)

The Gigabyte GA-7TESM was the first X58 board to be able to use 16Gb sticks, but unsure whether or not you can use 8Gb sticks with your board..

 

[Makaveli]

Doesn't matter what you say..They're different.. Intel even had different product codes for the X5600 series.. Any time you want to post your best stable cinebench scores. I'll be happy to school you further.. They were expensive for a reason.

"two product codes are used, 80613 for the UP desktop/server models and 80614 for the Xeon 5600-series DP server models."

http://en.wikipedia.org/wiki/Gulftown

lmao you are a broken record.

Your 24/7 non stable 4.8 ghz cinebench scores don't mean anything to me.

or those useless passmark benchmarks you were trying to use before.

They were expensive for reasons you have no clue about and that is obvious.

I would expect more logic from someone of your age. School me that is hilarious i've been on this forum a heck of alot longer than you.

 

[Burpo]

What does it matter how long you've been on a forum? lol THAT's HILARIOUS!
This was my first Xeon in 1998.. You?

I'll leave you to your thoughts/excuses & stick with intel's FACTS! The X5600 series are different from the W series & desktop i7's.. as denoted by INTEL..

 

[Makaveli]

What does it matter how long you've been on a forum? lol THAT's HILARIOUS!
This was my first Xeon in 1998..

I'll leave you to your thoughts/excuses & stick with intel's FACTS!

You don't even know that second QPI link goes unsued in the single socket boards that you are guys are throwing these xeon's in and was trying to use that a point to further your argument and I have excuses right.

Dude save it.

You obviously have alot of free time on your hands and I think its because your retired. But posting these xeons are better overclockers is incorrect. When you compare 32nm to 32nm the results are similar why because its the same arch.

You post in this thread have been helpful to many an appreciated, but if you continue to post stuff that is incorrect I will correct you.

 

[Burpo]

lol.. you're gonna correct intel.. LMAO..
They're different! INTEL says so! http://en.wikipedia.org/wiki/Gulftown
Yes, I have always known the 2nd QPI link goes unused.. Have been playing with Xeons & server procs a LONG time..

 

[Makaveli]

lol.. you're gonna correct intel.. LMAO..
They're different! INTEL says so!
Yes, I have always known the 2nd QPI link goes unused.. Have been playing with Xeons & server procs a LONG time..

I'm not correcting intel i'm correcting you.

SHOW ME PROOF that these 32nm 95watt xeon's overclock better than a 32nm 130watt i7.

And a 5Ghz suicide benchmark screenshot doesn't cut it.

i've been following this thread since day one and nobodys numbers in this thread support that claim.

 

[Burpo]

You're not correcting me.. I'm at 4400 stable, and use it regularly for rendering.. Please show me your i7 at that speed under load with those temps.

 

[Makaveli]

You're not correcting me.. I'm at 4400 stable, and use it regularly for rendering..

And is that vcore correct in the CPU Z shot ??

I know sometimes it can be off by abit!

 

[Burpo]

Sure is.. as I've mentioned before, I have no problem with voltage peaks up to 1.5v, as long as you control temps..

Tech support conversation with intel..
"Intel: Please wait for a site operator to respond. You are now chatting with 'Daniel'
Guest: TJ max , Tcase for 3960x
Daniel: Hello. Thank you for using the Intel Customer Chat Support service. We are glad to be of service. How may I help you?
Guest: Can you explain to me the temperature limits for my processor ?
Daniel: I can tell the normal operating temperature is 66.8°C and can go up to 100 degrees and eventually will shut down to prevent damage to the processor
Guest: At what temperature do i start degrading my processor ?
Daniel: there is no such temperature, processors do not degrade since they are designed to work at high temperatures."

http://www.tomshardware.com/forum/353195-28-tcase-processor-degradation

 

[Makaveli]

Sure is.. as I've mentioned before, I have no problem with voltage peaks up to 1.5v, as long as you control temps..

Whoever that daniel guy is he is wrong.

And there are numerous people on this forum alone that have had cpu's degrade from overclocking they have had to increase voltage to keep the same clocks or after the while the processor will only run at default clocks i've seen this first hand. I would bet a full paycheck that a processor that is run at 95c for 5 years straight compared to one that ran at 70c for 5 years which show better longevity.

When I get home i'm going to up my clocks to 4.4 and do a render or run OCCT for a few hours and post back what vcore I needed to keep it stable.

 

[Venoom]

Guys, please stop fighting

SHOW ME PROOF that these 32nm 95watt xeon's overclock better than a 32nm 130watt i7.

And a 5Ghz suicide benchmark screenshot doesn't cut it.

would you be convinced if I can (or fail to) get my X5675 (3.06GHz) to 4.5GHz, stable and everything, as opposed to your core i7-970 (3.2GHz 135W) that you got to 4.2GHz?

What can I do to settle this debate?

 

[Burpo]

Sure, high voltage will degrade a CPU (slightly) if kept at constant high voltage. That's the problem with your board/bios. You don't have adaptive voltage. A board that does will idle at 1.18 volts and peak much higher when needed. Under these conditions, the CPU will degrade MUCH slower.. Intel says I can run my Xeon at 96c peak for 10+ yrs.. I believe them.

 

[Makaveli]

Guys, please stop fighting



would you be convinced if I can (or fail to) get my X5675 (3.06GHz) to 4.5GHz, stable and everything, as opposed to your core i7-970 (3.2GHz 135W) that you got to 4.2GHz?

What can I do to settle this debate?

No argument just a difference in opinons.

You will be able to hit 4.5 but the question is at what vcore and will it be 24/7.

I can also up my clocks too but am I willing to put the vcore into the chip for it to hold. All of my overclocks are 24/7 and it has to be stable I can also post a 4.8Ghz screenshot running for 2 mins but what is the point of that.

My whole point is all the 32nm chips are the same arch yes the xeon's will have more qpi etc and more testing and validation done for the environments they run in but don't get that confused with better overclocking product. The overclocking headroom in all the 32nm chips are similar.

 

[Makaveli]

Sure, high voltage will degrade a CPU (slightly) if kept at constant high voltage. That's the problem with your board/bios. You don't have adaptive voltage. A board that does will idle at 1.18 volts and peak much higher when needed. Under these conditions, the CPU will degrade MUCH slower..

in that adaptive overclocking setting the peaks will be far higher than my non adaptive overclock so I think it will degrade faster not slower you were peaking at almost 1.45 just to hit 4.4Ghz.

where as when I get home i'm pretty sure I can hit 4.4Ghz at 1.35v.

So yes your idle voltage will be lower with the adaptive I will check the idle part on my rig tonight. I know have most of power settings and C state turned on but I will double check.

 

[Venoom]

No argument just a difference in opinons.

You will be able to hit 4.5 but the question is at what vcore and will it be 24/7.

I can also up my clocks too but am I willing to put the vcore into the chip for it to hold. All of my overclocks are 24/7 and it has to be stable I can also post a 4.8Ghz screenshot running for 2 mins but what is the point of that.

if you want, you can post your settings, and I can see if I can get it to 4.5GHz with similar or better settings
stable of course, I never turn off my PC anyway.

I personally have no opinion, just bought the X5675 for the higher clock and lower TDP, and hoping it overclocks better than the 130W Xeons, but that remains to be seen.

 

[Makaveli]

if you want, you can post your settings, and I can see if I can get it to 4.5GHz with similar or better settings
stable of course, I never turn off my PC anyway.

I personally have no opinion, just bought the X5675 for the higher clock and lower TDP, and hoping it overclocks better than the 130W Xeons, but that remains to be seen.

yup no problem I will take a couple pics of the bios settings when I'm done work.

 

[Burpo]

The real test would be with a kill a watt meter.. Performance Per Watt = Winner

Guess I could go buy one..

 

[Burpo]

No argument just a difference in opinons.

The overclocking headroom in all the 32nm chips are similar.

Umm ok.. so my 2.66Ghz Xeon is overclocked by 70% to 4.5Ghz and your 3.2Ghz i7 is similar? 70% overclock would be 5.44Ghz..

Hmm, which overclocks better/has more headroom? 95 watt 2.66Ghz Xeon or 130 watt 3.2Ghz i7?

It is obvious to me, and can be deduced simply with a calculator, which one is stretching the most, and overclocking better..

Staples nor Radioshack near me had a Kill A Watt meter.. Do I really need one? It should be obvious..

 

[Venoom]

Interesting finding, my mild 3.44GHz overclock (all I did was bump the BCLK to 150, did not bump any voltage) is actually not stable: I got a blue screen as soon as I ran the Intel Burn Test.
Didn't get any blue screen before that during any gaming session. The test is obviously way more intensive on the CPU.

so I went back to stock speed (3.07GHz) and no more blue screen.

temps never went above 50°C.
without any OC, min temps are between 20°C and 25°C, and I am idling between 24°C and 28°C.
what a huge improvement over the 45nm 130W i7-920.

 

[indio22]

Part of the rub on switching from x58 to z97 platform, is the cost of the new CPU and board. But I got a sweet Microcenter combo deal last weekend, on an i7-4790K CPU and decent z97 board, for $300 total. If the Xeon costs me $100, that is only around a $200 difference. And I could also sell the X58 board and CPU to bring that difference down further. So for the most part, the cost saving sticking with x58 and getting Xeon is reduced in my particular case. For me the decision is more about performance and capability (PC is not used for games, mainly application design and video editing/effects).

z97 gets me proper fast Sata III for my fast Samsung SSDs (my x58 board has poorly implemented slow add-on Marvell Sata III), and gets me some on board USB 3.0 connectors and a few other things I would like. And i7-4790K is no slug either. I feel it represents a decent upgrade in performance and capability from the current setup. Maybe not as much an upgrade were I to install a 6 core Xeon, but still an upgrade in a number of respects (particularly in Sata III ports and single threading). Would you guys recommend sticking with my plan to swap the i7-4790k/z97 in place of the i7-950/x58? Or would you recommend sticking with x58 and getting a Xeon? (Keeping in mind the financial incentive to stick with x58 is a bit less in my case.)