X58 + i7 930 adventures

n7

Elite Member
Jan 4, 2004
21,281
4
81
n7 made some foolish impulse buys over the last couple weeks, & the result is pictured ahead. ^_^

Gigabyte X58A-UD3R
Intel i7 930
HIS HD 5870
Corsair H50
A bunch of other stuff too, but less exciting...

I've been running a Gigabyte P55-UD3R + originally an i5 750 (4.15 GHz) & more recently, i7 860 (3.9 GHz).
While i was extremely happy with that, i was getting a bit bored with it...

Some pics of the new toys.

Gigabyte&


overview.jpg


io%20panel.jpg


While the SATA3/USB3 is largely useless to me presently, the 10 SATA ports + 2 eSATA is delightful, as i run a LOT of HDDs over various systems.

I switched the H50 fan out for a couple beloved Noctuas (exhausting out).

push-pull%20exhaust.jpg


push-pull%20exhaust2.jpg


Batch is B: 3938B268

batch.jpg


I quickly discovered X58 is annoyingly more difficult to OC than P55, particularly regarding higher speed RAM, due to the precise calculations needed surrounding QPI speed + uncore speed + RAM speed.

I have a lot more to learn, & will work on lowering voltages, etc. more as i have time.

Below is the latest results i've gotten.
HT is indeed off, as i've found i do absolutely nothing that benefits from it.

4.3 GHz (HT-off) 1.3v BIOS w/ LLC level 2 enabled

21x205%20(4305%20MHz)%201.30000v%20llc2%201.315vtt%201.22qpi-pll%201.18ioh-core%201.62vdimm%2036x-qpi%2017x-ioh%203ch%20DDR3-1640%208-8-7-24-2N%20LinX%2025%20passes%20STABLE.JPG
 
Last edited:

aigomorla

CPU, Cases&Cooling Mod PC Gaming Mod Elite Member
Super Moderator
Sep 28, 2005
21,087
3,598
126
so you liking that 930? :)

N7 also if u can get another fan on that radiator on the other side, you could probably lower your temps by roughly 3-8C depending on how much air you can feed that radiator.

from your temps, unless your ambients is up there, its being starved.
 

Rubycon

Madame President
Aug 10, 2005
17,768
485
126
so you liking that 930? :)

N7 also if u can get another fan on that radiator on the other side, you could probably lower your temps by roughly 3-8C depending on how much air you can feed that radiator.

from your temps, unless your ambients is up there, its being starved.


Probably not too bad. I was around 90C with the H50 with a pair of GT 1800RPM fans in push pull on the radiator. W3520 at 4.4GHz, HT ON, 1.3V. Megahalems temps around 80 with same config. Ambient around 20-22C. The H50 isn't really for the "over 4GHz Quad crowd" as far as I'm concerned.

HOWEVER you must remember that Linpack puts temps at the upper limit far from anything you encounter with everyday computing. So unless you get your jollies off packing lins, I don't see THAT big of a deal. A system touching 90C in Linpack would probably be in the 60s when gaming, etc. Perhaps mid 70s when encoding. Unless your ambients are high and/or the radiator gets clogged with dust. (keep it clean!)
 

n7

Elite Member
Jan 4, 2004
21,281
4
81
I don't know how good my i7 930 is, but TBH...doesn't seem very good from what i've seen.

It's needing ~ 1.32v load w/o HT for 4.3 GHz range, which is not great considering many have 920s w/ HT on & less vcore @ the same or better speeds.

I guess it could be worse always though.

I forgot to mention, i did switch the fan on the H50.

I put on two Noctua P13s in push-pull exhaust on it.
LOL @ Corsair's recommendation of intake, no thanx.

I have a Haf 932 case, & honestly, airflow really isn't an issue in there.
The rad barely gets warm under load for the H50 from what i can feel.

I think i'm just really pushing its abilities w/ a hot chip.
I'm actually rather tempted to get the NH-D14, but i really don't like the idea of that much weight/size on the board @ an angle regardless of how excellent Noctuas mounting systems are.

And yeah, temps don't scare me @ all, as once i'm done stability testing, it'll never be even getting warmed up as UT3 is likely the most intensive thing it'll be running.

Thar we go, lovely beige babies. :awe:
push-pull%20exhaust.jpg


push-pull%20exhaust2.jpg
 
Last edited:

Rubycon

Madame President
Aug 10, 2005
17,768
485
126
Corsair's recommendation - case cooling bedamned - will net cooler CPU temps - ONLY! A HAF932 even stock runs very close to ambient all closed up so there would be little to gain (or in the case of temps - lose) by having the fan pull air in. As a matter of fact when you start bucking flows actual flow THROUGH the radiator can be LOWER!.

What about HT ON? You should run it ON to see what your max clocks look like. Don't be afraid to pour on the VCORE - all the way up to 1.4V. Some go higher but unless you want to start a fire you should NOT do this with an H50. ;)

VX/Mega Shadow temps are close to 10C cooler than the H50. I know I said it in the past but I WILL get results of what the H50 can do with its radiator immersed in a bucket of ice water! :biggrin: Perhaps the same time Aigo gets his multi TEC exchanger hooked up and RUNNING! :p

Pointless or perhaps not. What the hell does it show? Well if you increased the size of the radiator I suppose. I don't know it seems to me if someone goes to all the trouble of hacking up an H50 to put it back together again making things even more prone to failure as well as complex they might as well buy all the junk that goes along with a forbidden realm known as water cooling. :eek:
 

n7

Elite Member
Jan 4, 2004
21,281
4
81
Corsair's recommendation - case cooling bedamned - will net cooler CPU temps - ONLY! A HAF932 even stock runs very close to ambient all closed up so there would be little to gain (or in the case of temps - lose) by having the fan pull air in. As a matter of fact when you start bucking flows actual flow THROUGH the radiator can be LOWER!.

Well, if i get un-lazy i might try that.
You are right, case airflow is superb in the Haf 932 anyway.

What about HT ON? You should run it ON to see what your max clocks look like. Don't be afraid to pour on the VCORE - all the way up to 1.4V. Some go higher but unless you want to start a fire you should NOT do this with an H50. ;)

I'm not afraid of vcore, but the H50 won't be able to handle that anything close to that number w/ LinX :hmm:

I will determine max w/ HT eventually, but this is brand new, so it'll take me a bit of playing time before all the answers are revealed.

The main problem w/ HT for me is that i found going from the no HT of the i5 750 to the HT on w/ the i7 860 gained me nothing but higher temps.
I don't do anything that utilizes 8 threads, so for gaming, the higher clocks are preferable to lower clocks w/ more threads...

VX/Mega Shadow temps are close to 10C cooler than the H50. I know I said it in the past but I WILL get results of what the H50 can do with its radiator immersed in a bucket of ice water! :biggrin: Perhaps the same time Aigo gets his multi TEC exchanger hooked up and RUNNING! :p

Pointless or perhaps not. What the hell does it show? Well if you increased the size of the radiator I suppose. I don't know it seems to me if someone goes to all the trouble of hacking up an H50 to put it back together again making things even more prone to failure as well as complex they might as well buy all the junk that goes along with a forbidden realm known as water cooling. :eek:

Sounds like fun; do share the results with us. :D
 

Rubycon

Madame President
Aug 10, 2005
17,768
485
126
I know how much Aigo loves those Noctuas - so much he calls 'em Noctunas! :D

Was there a decent amount of air coming out?
I got another 3C off by going to (2) Delta 155CFM fans but I feared I'd have to get an unlisted head to stop the ring-ring-rrrringing in my ears just after a few rounds of ssssuper-pi. (cake is better but does miserable between the IHS and bottom of the cooler!)

Got an IR thermometer handy? It's good to record radiator temps, etc.

It's all about the numbers. ;)
 

n7

Elite Member
Jan 4, 2004
21,281
4
81
IR Thermometer, nope.

I can only imagine all the fun gadgets you have around. :sneaky:

TBH, i'm pretty picky about noise.
While i can tolerate the louder whooshing noise from the big fans on the Haf 932 & don't mind the GPU getting louder in games (i use headphones for gaming), anything more wouldn't do for me.

That means poorer temps & less insane OCs, but i've been quite happy with my "weak" Noctuas.
The OCs i've gotten using them on well, pretty much every CPU cooler i've ever owned, have been adequate for my needs. :hmm:
 

Apocalypse23

Golden Member
Jul 14, 2003
1,467
1
0
A beautiful mobo n7! I've already got my eyes on this one, will buy one as soon as my 920 and 6gig ddr3 1600 ram arrive, cheers!

PS-I'll be asking you for some help to set mine up with an OC :)
 
Last edited:
Nov 26, 2005
15,194
403
126
Could running 4 sticks of ram be causing a problem with your overclocking?

I hear if you take a picture of a shock rifle and wave it around in front of the ram it might settle the ram down and produce better overclocking results, but, meh, it's just what I heard.

EDIT: wait, did you say UT3? I think once I get my game rig up and running I'll install UT2k4.... as of right now, i still have to install BF:BC2, Avatar online, & AvP
 
Last edited:

n7

Elite Member
Jan 4, 2004
21,281
4
81
Could running 4 sticks of ram be causing a problem with your overclocking?

I hear if you take a picture of a shock rifle and wave it around in front of the ram it might settle the ram down and produce better overclocking results, but, meh, it's just what I heard.

EDIT: wait, did you say UT3? I think once I get my game rig up and running I'll install UT2k4.... as of right now, i still have to install BF:BC2, Avatar online, & AvP

Wait, you finally bought UT2k4? :awe:

Yay, now you can play the less grounded-faster-paced-better-version of UT w/ me :D

I don't really play UT3 anymore, but it is technically still the most CPU-intensive thing i'd be doing once i stop stress testing (if i ever played it anyway) :hmm:

I highly doubt 4 DIMMs is hurting anything, as what's limiting me presently is temps, nothing else...
I fully plan to roll w/ the 4 DIMMS so i can slam screenies in the face of any n00b i see stating you need 3 or 6 DIMMs for X58 & how dual channel isn't an option, etc...

BTW, stock speed dual channel (slots 1+2+3+4) vs. triple channel (slots 1+3+5+2) synthetics!

Triple channel is a teeny bit faster for me, but not by much.

dual channel (slots 1+2+3+4)
i7&


triple channel (slots 1+3+5+2)
Triple%20channel%20stock%20Everest-MaxxMem-SuperPi.JPG

Disregard the CPU speeds; Speedstep was doing funny things as usual, in those screenshots.
 

Anubis

No Lifer
Aug 31, 2001
78,712
427
126
tbqhwy.com
A beautiful mobo n7! I've already got my eyes on this one, will buy one as soon as my 920 and 6gig ddr3 1600 ram arrive, cheers!

PS-I'll be asking you for some help to set mine up with an OC :)

im running the same mobo, I also agree its pretty nice, my 920 is running 20x200 with HT on, cant really complain
 

Gillbot

Lifer
Jan 11, 2001
28,830
17
81
IR Thermometer, nope.

I can only imagine all the fun gadgets you have around. :sneaky:

TBH, i'm pretty picky about noise.
While i can tolerate the louder whooshing noise from the big fans on the Haf 932 & don't mind the GPU getting louder in games (i use headphones for gaming), anything more wouldn't do for me.

That means poorer temps & less insane OCs, but i've been quite happy with my "weak" Noctuas.
The OCs i've gotten using them on well, pretty much every CPU cooler i've ever owned, have been adequate for my needs. :hmm:

JamesIR.jpg


If you want a cheap-o IR Thermometer, get one from here:
http://www.harborfreight.com/cpi/ctaf/displayitem.taf?Itemnumber=93984

If you wait for the sales, you can get this one for $25 or less with coupons and etc. http://www.harborfreight.com/cpi/ctaf/displayitem.taf?Itemnumber=96451
 

jtisgeek

Senior member
Jan 26, 2010
295
0
0
You need to invest in a real water cooling setup if your ever going to get up there even more with good temps. With the money you have already spend shouldn't be to bad if you can swing it .

Invest in a good setup you can use for years to come and your be happy.
 

Rubycon

Madame President
Aug 10, 2005
17,768
485
126
If you want a cheap-o IR Thermometer, get one from here:
http://www.harborfreight.com/cpi/ctaf/displayitem.taf?Itemnumber=93984

If you wait for the sales, you can get this one for $25 or less with coupons and etc. http://www.harborfreight.com/cpi/ctaf/displayitem.taf?Itemnumber=96451

Problem with the cheapies is sub 10:1 and (most importantly!) fixed emissivity! This is important if you shoot metal parts, for example. (I'm sure you know this if you're using LIR. ;) )

There are ones with adjustable emissivity for under $100 I'm told. Fortunately in the electronics field everything that's shot is below the charring point of masking tape. What we do is put a thumb sized piece on something we want to measure, come back in 15 minutes and put the laser on the tape. The reading is recorded. Then the laser is pointed on the object right next to the tape. E is adjusted until the temp reading matches when we shot the tape. Works great and no chart needed or guessing material type which can easily be wrong if its coated or you just have no idea what it really is. ;)
 

Gillbot

Lifer
Jan 11, 2001
28,830
17
81
Problem with the cheapies is sub 10:1 and (most importantly!) fixed emissivity! This is important if you shoot metal parts, for example. (I'm sure you know this if you're using LIR. ;) )

There are ones with adjustable emissivity for under $100 I'm told. Fortunately in the electronics field everything that's shot is below the charring point of masking tape. What we do is put a thumb sized piece on something we want to measure, come back in 15 minutes and put the laser on the tape. The reading is recorded. Then the laser is pointed on the object right next to the tape. E is adjusted until the temp reading matches when we shot the tape. Works great and no chart needed or guessing material type which can easily be wrong if its coated or you just have no idea what it really is. ;)

Still pointless, even with adjustable emissivity the readings are nowhere near accurate with any material below ~0.7 due to reflectivity, etc.

Also, spot size is basically irrelevant anymore. If you can get close enough with the proper focal range, spot size has next to zero influence on the temp readings as long as the item you are monitoring isn't "too small"/.

TBH, when I shoot with the camera I never adjust emissivity anyway. There is no point because as I said above, anything below ~0.7 emissivity isn't an accurate reading anyway. You just need to make note of it, take the proper steps to increase the emissivity to a higher level or use the proper CONTACT temperature measuring equipment if you need a true value.
 

Rubycon

Madame President
Aug 10, 2005
17,768
485
126
I've verified lower than 0.7 readings with contact method as within instrument tolerance. Now if you're shooting mirror finish...
 

Rubycon

Madame President
Aug 10, 2005
17,768
485
126
How many reefers (people that keep reef aquaria!) with multiple kW lighting setups, chillers, $10,000 monthly power bills, and windows open in the winter - you name it - get tagged as pot growers? :eek:
 
Nov 26, 2005
15,194
403
126
How many reefers (people that keep reef aquaria!) with multiple kW lighting setups, chillers, $10,000 monthly power bills, and windows open in the winter - you name it - get tagged as pot growers? :eek:

is this a trick question?

i don't know... alot?
 

n7

Elite Member
Jan 4, 2004
21,281
4
81
Did a bit more fine tuning today.
Lowered some voltages, using LLC1 instead of LLC2; vdrop & vdroop are annoying, but consistent so i'll likely leave as is.

Couldn't get HT-on even stable @ 4.15 GHz without more voltage than my tolerance for temps allowed, so don't think i'm going to bother worrying about HT until something i start running in the future benefit from the extra 4 threads.

i7&



Unrelated but here's my HDD synthetics.
I realize the writes are a bit low, but that's likely due to them never seeing TRIM action to fix them from their previous broken state (as previously i was running Vista, not 7, & haven't secure erased them).

This is 2x X25-M 80 GB G2 + one X25-M 80 GB G1 in RAID-0 on Intel ICH10R
03.26.10%20triRAID%20benching.JPG
 
Last edited: