ASUS Striker II Extreme

[Leeghoofd]

That's already a nice result with the Striker man... Quads on Nvidia is a pain in the b u t t ... with DDR3 it's even harder

[Trigger5521]

Gave up on the QX9650, it seems to dislike high FSB with 4 cores enabled...

Still chasing the Holy Grail of 500 MHz FSB on a quadcore, it's time to move onto the Xeon X3380...

X3380 ALL 4 CORES - FSB 491 MHz (1964 MHz QDR) w/8GB RAM





X3380 491 MHz FSB P1 & P2 Auto (Poor chipset latency & bandwidth)



Close but no cigar.

Fails to boot into Windows above 1964 (will POST though).

BIOS settings for 491 MHz FSB;

1964:1637 (FSB:RAM)
Vcore 1.45V LLC enabled (1.488V)
PLL 1.60V (1.64V)
VTT 1.40V (1.45V)
VDIMM 1.64V (1.71V)
NB 1.58V (1.60V)
SB 1.60V (1.58V)
NB GTL -20mV
CPU GTL 0 -95mV
CPU GTL 1 -56mV
CPU GTL 2 -56mV
CPU GTL 3 -95mV
P1 & P2 Auto

Give the RAM more juice! 8GB needs 2.0v for me.

[chinobino]

Give the RAM more juice! 8GB needs 2.0v for me.

This RAM is rated for 1.5v @ DDR3-1333 or 1.65v @ DDR3-2400 (XMP profile), I'm already giving it 1.71v.

Unfortunately the NB doesn't want to run high FSB with RAM 2:1, even though the RAM is more than capable of it.

[Trigger5521]

This RAM is rated for 1.5v @ DDR3-1333 or 1.65v @ DDR3-2400 (XMP profile), I'm already giving it 1.71v.

Unfortunately the NB doesn't want to run high FSB with RAM 2:1, even though the RAM is more than capable of it.

All the ram I have tried wouldn't even post at 1.65v stock speeds. I can't remember I think Patriot is rated to 2.0v I'm actually running them at 2.06v for over a year stable no issues.

[slim142]

interesting chinobino

is that RAM made for tripple channel or dual channel?

I always wanted to know how much voltage would 1.65v tripple channel RAM need to run at let say 1800Mhz on the S2E. Same thing for dual channel RAM designed for lynnfield.

[chinobino]

The RAM is dual channel for 1156 boards.

This kit uses 1.65v right up to 2400 MHz, so 1800 MHz is easy (8-8-8-22 2T).

[slim142]

Is 1800Mhz 7-7-7-18 complicated as of now with this new ram kits? and what voltage for this?

I know 1.90 is the max safe to use on the striker, but I dont think you wanna use 1.90 on a 1.65 designed kit

[chinobino]

It depends more on the type of IC than anything.

I have PSC (Samsung?) IC's and they will do DDR3-1600 at 7-7-7-21 2T with 4GB (2 x 2GB sticks) but the NB can only handle DDR3-1800 at 8-8-8-22 2T with 8GB (4 x 2GB).

[slim142]

I see
but 1800 8-8-8-22 still not that bad.

What tRC?

[Sangrak]

All the ram I have tried wouldn't even post at 1.65v stock speeds. I can't remember I think Patriot is rated to 2.0v I'm actually running them at 2.06v for over a year stable no issues.

Yeah, the Patriots are rated for 2.0v. I've been running them at 2.00-.02 for about the same amount of time.

[chinobino]

I see
but 1800 8-8-8-22 still not that bad.

What tRC?

tRC is 43 (on Auto at the moment).

[slim142]

I like my tRC around 30 but at 1800 8-8-8-22 I dont think that would be possible. Maybe 38?

I noticed that gives a good boost in performance, but its just as important as the CAS and the main ones.

[chinobino]

Thanks Slim, good to know.

The tRC for DDR3-2400 is 42 by SPD (with 4GB) so I might have to try tightening that up a bit.

tRC (Bank Cycle Time) - [Auto, 1 ~ 63] This is arguably one of the most important timings when it comes to memory performance with NVIDIA chipsets. The rule when it comes to tRC is you should set it no lower than tCL + tRAS + 2. Even small reductions in this timing can bring about large decreases in memory access latencies.

[Edit] Some results...

Max stable 8GB RAM frequency with RAM:FSB ratio 2:1



tRC 46 (Auto) P1 & P2 Enabled





tRC 40 P1 & P2 Enabled





Thoughts?

[slim142]

You might have reached your max bandwidth thats my only guess. I think the highest Striker I have seen got up to 14k, but 12k-low 13k is most common.

Did you try 41, 42, 43, 44 and 45?

Also, just for the heck of it try 39 and 38 to see if it helps at all.

[chinobino]

I tried tRC all the way down to 16 and didn't notice much change in bandwidth (+50MB/s read).

I wonder if there is a better way to check tRC gains than using Everest's memory bandwidth test?

[slim142]

tRC down to 16? How did you do that? Usually below 30 it will not boot (on high OCs)

Are you using the latest version of everest? EVEREST Ultimate Edition v5.50.2242 Beta

http://www.lavalys.com/sites/default...build_2242.zip

[chinobino]

Well, here's some screenshot's;

tRC 16 P1 & P2 Enabled



Everest Ultimate v5.50.2242 Beta



Is there a better way to measure memory access latencies?

[slim142]

Ok that is just weird.

I dont know what to tell you. But at this point all you can do is try Sandra and see if that gives a different result so you can compare.

[StefanZ]

Hello All,

Can i flash my Asus Striker II NSE with the Asus Striker II extreme 1402 bios ??

[slim142]

I think people in the ASUS II NSE forum have done so and didnt have problems.

But that was back in the 701 - 901 days. Current BIOS is 1402 I cant guarantee everything will be fine, even though it should be

[chinobino]

Hello All,

Can i flash my Asus Striker II NSE with the Asus Striker II extreme 1402 bios ??

As Slim said - it should be fine.

I just wanted to share the following discovery;

If your CPU has a half multiplier by default (e.g. x3380 = 9.5, Q9550 = 8.5) and you set the multiplier to a whole number (i.e. 7 or 8 or 9) whilst leaving EIST (Enhanced Intel SpeedStep) enabled, then the CPU will go up to the next available half multiplier when Windows loads, overriding the multiplier set in BIOS.

But if you disable EIST then the multiplier stays limited to the BIOS set max multiplier (a whole number, in this example).

Using BIOS 1402.

[Bobly]

Okay it's that time of year again... When I need to install a big program, extract a big archive or move some big files again, and I start getting the dreaded freezes again...

I say this because my S2E is behaving very well, good a nice stable overclock, stock volts on NB, SB and DDR with just a little bit extra on the CPU, everything remaining nice and cool... And then it comes to moving big files...


Now I know what you're thinking, high I/O, pump up the SB? Will I've tried adding volts to the SB, the NB, the memory modules (not so much, don't think I've ever tried going above 1.92, 1.94, they're rated at 1.9), but nothing seems to make a difference...


So here I am again... Help me Even just a point in the right direction could be a godsend.


(Btw: Saw 1402 was out, figured it couldn't hurt, set BIOS to default before the flash and forget to re-enable the overclock straight away. Freeze on big file operations (extracting an 8GB rar in this case but I know this is the same issue) still happening. Other than that it passes overnight memtest, prime and intel burn >.> )

[slim142]

What windows are you using?

I havent had problems with files up to 2gbs big. I havent tried bigger files in a really long time tho. I would double check your RAM. That used to be my biggest problem back in the day when it came to corrupted files.

Also, you might want to reflash your BIOS with awdflash since I heard ezflash leaves corruption. Idk if this is a rumor or not, but doesnt hurt to try.

[Bobly]

OS: Windows 7 64

I used to have 1301 until last week and that had been flashed using awdflash, as had the one I was using before (this is 2 years ago now, not sure which one it was :P ) but the same problem always occurred.

You're probably right on the RAM front but I have no clue which setting is triggering these errors. I'll go and grab another copy of memtest and run that again but last I remember it was passing it fine.


This really is an irritating issue though, coz I tend to go a couple of months without dealing with big files and completely forget about it... Then try to extract a big archive, or install a big game, or backup a bunch of data... and the freezes start again :P

I've actually resorted to network mounting the archive on my EEE running Ubuntu to see if it'll work better if I extract it using that one... ridiculous having quad 3.8GHz cores idling while the poor atom is chugging away with unrar and samba :P

(For what it's worth, doing it over the network doesn't "fix" the issue, still crashed the first time, but on attempt #2 I've so far made it to 30% but going to have to wait an hour or so to see if it completes)



Edit: For what it's worth I think I've managed to narrow it down to the jMicron SATA controller.

I'm going to be heading off for the holidays so I wanted to move stuff over to the external and the constant crashes was making that next to impossible so at every reboot I would change from of the BIOS settings. Eventually I just disabled the jMicron controller and that didn't help, but when I re-enabled it and set it to IDE rather than AHCI it seems to have helped as it's been copying for 2h+ now and it's still going strong...

Now, am I going to have to run it in AHCI most of the time and switch to IDE whenever I need to do big transfers or can this somehow be fixed?

[fauka]

this is my score all working good ; )

i cant set fsb ratio 1:2 because my ram is a bit low ; P

edit i just make stable 7-7-7-20 1T