ASUS Striker Extreme 1301 BIOS 7/11 Update

[Leeghoofd]

get some decent skills + cooling

Help peeps then if you have THE skills in stead of posting comments which are pointless...and help noone...
Not everybody is gonna watercool their PWM area let alone just the NB, so running 500FSB on a rig on air is suicide , if you want to prove me wrong put all on air, go ahead we will talk again in half a year.
If you thoroughly tested the board you will notice that performance degrades rapidly when surpassing certain FSB and secondly my 632A doesn't do 500FSB it's limited around 485FSB

[markost]

Help peeps then if you have THE skills in stead of posting comments which are pointless...and help noone...
Not everybody is gonna watercool their PWM area let alone just the NB, so running 500FSB on a rig on air is suicide , if you want to prove me wrong put all on air, go ahead we will talk again in half a year.
If you thoroughly tested the board you will notice that performance degrades rapidly when surpassing certain FSB and secondly my 632A doesn't do 500FSB it's limited around 485FSB

Right again my friend

[Ambro]

Help peeps then if you have THE skills in stead of posting comments which are pointless...and help noone...
Not everybody is gonna watercool their PWM area let alone just the NB, so running 500FSB on a rig on air is suicide , if you want to prove me wrong put all on air, go ahead we will talk again in half a year.

My point was to get some descent cooling if you want to run a high FSB - this IS xtremesystems. I'll let you know how this board is going in 6 months @ 500FSB, unfortunately the water stays

My sig contains enough info for people to follow what I've done - if they can't read between the lines then, yes, they might think it was a useless comment.

Faiakes has shown he had the skills to grab some of the highest benched scores with his ram setup as he detailed - as also my sig details

If you thoroughly tested the board you will notice that performance degrades rapidly when surpassing certain FSB and secondly my 632A doesn't do 500FSB it's limited around 485FSB

I think the previous post to yours by Faiakes shows there is a difference above 424FSB with sync'd ram speeds. Are you suggesting he, like myself, is not thoroughly testing this board.

For the record - in my experience performance IS gained, although not by a great amount ... once again - this IS xtremesystems; so any little extra is what I am here for.

Right again my friend

Abrasive my previous post may have been to you too; it may have given others the resolve to try for 500FSB to milk their systems before testing pmp's new bios. If you're content to sit and wait then thats up to you.

Don't argue with a fool, the spectators can't tell the difference

[markost]

Abrasive my previous post may have been to you too; it may have given others the resolve to try for 500FSB to milk their systems before testing pmp's new bios. If you're content to sit and wait then thats up to you.

Don't argue with a fool, the spectators can't tell the difference [/QUOTE]

Unfortunately experience can be misinterpreted by some as narrow mindedness or not being seen to be willing to push the envelope. Xtreme Systems generally is about total system performance and not just extreme FSB. There are other threads which are dedicated solely to that topic. It is a known fact that complete system performance drops substantially on the Striker after 424 FSB. If a user is solely content with achieving the highest possible FSB speed at the expense of overall system performance then that is his or her prerogative.
Having built and tuned multiple machines utilizing Strikers with varied associated hardware using Air, Water and Phase cooling, none of those machines have benefited from FSB speeds over 424. All that is being said here is that if you want the best performance from this board, at this stage, with current BIOS revisions, it is more productive to tune it's memory performance rather than go for the big FSB numbers. I've had numerous Striker's all with late BIOS versions stable at 500 FSB but have always dropped back to lower than 425 and gained much improved system responsiveness.
I and a few others here, are simply stating what the limitations of this board are (currently) and trying to assist other inquiring Striker users into gaining the "highest level of performance" and not just "the highest FSB numbers".

The choice is of course, yours.

And you are correct, I will not argue with a fool. Some spectators can tell the difference

[Leeghoofd]

Well my point is and stays and it's not my point to start a flaming war here :

On air (let's say to run the board out of the box) it ain't guaranteed to keep those clocks, I've had numerous discussions about peeps ing mobo/ram manufacturers that the prodcts they got were crap , they ran their rigs up to full potential with dangerous volts on air, finally the mobo, ram or both died and then they start complaining it's a crap mobo...

I'm indeed a safe overclocker , only pushing the rig when having fun or testing some new ram and co , I just want to spare some users the headaches & sleepless nites that not all mobo's or other hardware can't be pushed so hard as some members here...

I envy your rig as my board has never released it's full potential ( CPU limited here, as the cpu's do exactly the same on the P5K DLX ) but I don't even see, feel a difference at my games between 3.4 or 3.8 GHz as my TFT supports...max 1280 x 1024...

Extremesystems has changed more into a general forum where overclocks still lead the pack but more and more users tune in for general help then asking help for pushing the envelope, me being an very active rage3d and Abxzone forum member I'm still trying to aid those that need help , based on my own experiences ofcourse and that of other users...

Kinda interested in ya cooling setup Ambro , do you have a link where I can see more piccies plz as I might order too the full EK setup and do a second loop with smaller tubing for NB and PWM and keep the P5K Dlx on air... and go flat out on the Striker...

Cya lads , glad to see some users are at least happy wiht our superexpensive Strikers

[Ambro]

It is a known fact that complete system performance drops substantially on the Striker after 424 FSB.

Simple - are Faiakes results better @ a high FSB clock ( above your magic 424FSB figure)

YES

improved system responsiveness @ lower FSB's - I don't believe; especially when the 3X bench programs of Faiakes (and mine I should add) say otherwise

Who are all these others to back you up about your 424FSB ceiling that you refer to??

Well my point is and stays and it's not my point to start a flaming war here :

On air (let's say to run the board out of the box) it ain't guaranteed to keep those clocks, I've had numerous discussions about peeps ing mobo/ram manufacturers that the prodcts they got were crap , they ran their rigs up to full potential with dangerous volts on air, finally the mobo, ram or both died and then they start complaining it's a crap mobo...

I'm indeed a safe overclocker , only pushing the rig when having fun or testing some new ram and co , I just want to spare some users the headaches & sleepless nites that not all mobo's or other hardware can't be pushed so hard as some members here...

I envy your rig as my board has never released it's full potential ( CPU limited here, as the cpu's do exactly the same on the P5K DLX ) but I don't even see, feel a difference at my games between 3.4 or 3.8 GHz as my TFT supports...max 1280 x 1024...

Extremesystems has changed more into a general forum where overclocks still lead the pack but more and more users tune in for general help then asking help for pushing the envelope, me being an very active rage3d and Abxzone forum member I'm still trying to aid those that need help , based on my own experiences ofcourse and that of other users...

Kinda interested in ya cooling setup Ambro , do you have a link where I can see more piccies plz as I might order too the full EK setup and do a second loop with smaller tubing for NB and PWM and keep the P5K Dlx on air... and go flat out on the Striker...

Cya lads , glad to see some users are at least happy wiht our superexpensive Strikers

what can I say - your point had to be made and it is indeed true and correct

my volts are pretty mild with these clocks for MY system; stock boards may have issues.

Safe overclocking? Yes but as I said before this is XS
(see Attached Image )

(I should add this clause to all my future posts )

I am confident that a 500FSB is O.K - but only time will tell. 6 months we shall all know.

I run 1920X1200 - maybe it is the res but I do notice the difference when gaming on a good FSB clock. Lost Planets is especially responsive @ higher clocks which gives me hope when DX10 rules supreme.

My GTX runs 650gpu (I think normally 575) and 1070mem (I think normally 850) thanks to eddy's blocks.

I will post some picks soon in the liquid cooling area, thanks for the interest; other ppls machines inspired me initially - I only hope to do the same because this should be a fun hobby where users get something out of this forum

[Leeghoofd]

I run 1920X1200 - maybe it is the res but I do notice the difference when gaming on a good FSB clock. Lost Planets is especially responsive @ higher clocks which gives me hope when DX10 rules supreme.

My GTX runs 650gpu (I think normally 575) and 1070mem (I think normally 850) thanks to eddy's blocks.

If we can believe the guys at hardocp (the guys that knackered completely the Striker because they couldn't get a decent overclock out of it) even our 8800GTX sucks when running DirectX10 at it's full glory...

http://enthusiast.hardocp.com/articl...50aHVzaWFzdA==

[markost]

Well unfortunately there comes a time which someone dreads. To cough up and admit that one was rrrrrrrrong, well sort of. I just installed my first G0 chip into a Striker (E6850) and it behaves nothing like 1030 FSB chips regardless of which multi is used.
Initially I used an E6400, E6600, E6700, and an X6800 and they all exhibited the "crappy over 424 FSB phenomenon". With these chips, Everest Read would have been down 1000MB/s and latency up about 2.0ns. The exact opposite of the benches shown and with looser timings as well. Now it's time to take this chip off water and throw it on phase and see what happens.

[Faiakes]

Could you bring the mutli down to x8 and the memory to 4-4-4-12 so I can do a comparative test and confirm the results?

Aren't the final results affect by the CPU speed?

[Ambro]

Well unfortunately there comes a time which someone dreads. To cough up and admit that one was rrrrrrrrong, well sort of. I just installed my first G0 chip into a Striker (E6850) and it behaves nothing like 1030 FSB chips regardless of which multi is used.
Initially I used an E6400, E6600, E6700, and an X6800 and they all exhibited the "crappy over 424 FSB phenomenon". With these chips, Everest Read would have been down 1000MB/s and latency up about 2.0ns. The exact opposite of the benches shown and with looser timings as well. Now it's time to take this chip off water and throw it on phase and see what happens.

- time to start testing again, these G0 chips do behave differently (wish I had phase and a QX6850; that's a striker test waiting .....)

[Ambro]

If we can believe the guys at hardocp (the guys that knackered completely the Striker because they couldn't get a decent overclock out of it) even our 8800GTX sucks when running DirectX10 at it's full glory...

http://enthusiast.hardocp.com/articl...50aHVzaWFzdA==

lost planets is pretty but I think crysis will truely show off high end equipment. - with vista64; hopefully I'll have my other gtx back in time for that so I can SLI it. In a years time a gtx will be cannon fodder, but them there are the breaks if you want to keep up with the bleeding edge.

[slim142]

Markost

424Fsb we know that is better than having it at 430, 450 or even 490. However, I thought the good FSB was 425 and not 424?

[markost]

Markost

424Fsb we know that is better than having it at 430, 450 or even 490. However, I thought the good FSB was 425 and not 424?

Hey Slim
I recon that may change from board to board. On my board, 425 would drop performance, yet on a mates board using the exact same hardware it would do it at 426. On that note, with the E6850, I found a sweet spot of 1786 Mhz (4.01 Gig on water___putting on phase over the weekend) but on every reboot Everest latency would rise 2.5ns and memory read would fall 200 MB/s. I then dropped the FSB to 1785 Mhz and it stayed at the faster speeds every time. This is of course in unlinked mode, so it seems as if the speed holes on these Strikers may vary slightly from board to board (and also BIOS to BIOS)

[NightRaven]

but i would say that the variation isnt all too big and holes will always be there. we just got to test them out and see whats best

[markost]

but i would say that the variation isnt all too big and holes will always be there. we just got to test them out and see whats best

That's exactly correct

[ShoNuff]

No m8. It's not. Currently next major upgrade is still under qtc, just be patient
More details by beginning of next week, when i finish my vacation time.

Looking forward to the next bios...1401? While the 1303 is good it has room to grow when used with the Quads. Can't run my memory at tight with my Quad Core.

ex. Using 1303 in both examples...
With Dual core: 3-3-3-9-1T @ 400MHz.
With Quad core: 4-4-4-12-2T @ 400MHz (1T is completely unstable)

The board runs stable up to 410FSB for me. Hopefully 1401 will allow the nice FSB clocks and tight timings. If so, we will have a winner on our hands.

Thank you sharing your results with us here at XS.

[Malk4vi4n]

1303 fixed some coldboot problems my system was having. Anyways, I use 450FSB daily and got 475FSB 8hrs orthos out of the box, no voltage ajustments of any kind.

Indeed, I done this fsb testing with the old bios, didn't retest it after flashing.

[Leeghoofd]

I got that also with my E6400, just the Vcore needed to be changed, with the E6600 it required more voltage adjustements to keep it stable ( especially CPU VTT) probably due to the larger CPU cache and the extra strain on the memory controller... you got a good one mate keep it running like this for along time !!!

[Malk4vi4n]

I got that also with my E6400, just the Vcore needed to be changed, with the E6600 it required more voltage adjustements to keep it stable ( especially CPU VTT) probably due to the larger CPU cache and the extra strain on the memory controller... you got a good one mate keep it running like this for along time !!!

Yeah, exactly. I can make 3.8GHz almost stable @ 1.52v. But hey..that's not 24/7 voltages. I keep mine @ 1.46v and it run cool enough for winter. In summer i'll set it to 3.2GHz (1.3125v in bios and it's rock stable).

[markost]

These straps just keep getting stranger all the time. just threw in some Reaper PC2 9200 and bandwidth is definitely not uniform. When running memtest with memory set at 1t @ 1015Mhz it shows memory bandwidth as 5615 MB/s. Up that to 1025 Mhz and bandwidth jumps to 5815 MB/s. Fair enough one would think. I then raised memory speed to 1035 Mhz, then 1045 Mhz and finally 1056 Mhz all at 1t and 4.4.4.4 with tight subs. Bandwidth stayed at 5815 Mhz on each occasion. Everest Benches confirmed that at all these settings, speed did not change and remained practically identical to the screenshot. Anyone got any theories here ?? 1301 BIOS.

[markost]

Another question

Which is faster in real terms ??

Pic #1. 446 x 9 (unlinked) which has lower bandwidth & lower latency
Pic #2. 497 x 8 (linked & sync mode) which has higher bandwidth & higher latency

3D05 @ 446 x 9 = 21,316
3D05 @ 497 x 8 = 21,451

super PI - 1M @ 446 x 9 = 12.591
super PI - 1M @ 497 x 8 = 12.625

Super PI is slower at 497 FSB but 3D05 is higher even at slightly lower CPU speed and lower memory speed.

Lower FSB and higher memory speed (Unlinked) gives faster Super PI times, better memory latency, lower memory bandwidth
Higher FSB and lower memory speed (linked & sync mode) gives better 3D05 scores higher latency and higher memory bandwidth.

Which is faster?? In Windows, lower FSB and higher memory speed feels faster. using benchmarks alone.....who knows !

My call is that memory latency is the largest factor in speed but higher FSB certainly looks good if you like looking at numbers.

Worthwhile noting that 497 x 8 (unlinked) would not even post. I tested right up 515 FSB (linked) and it just got slower on all benches.

It would have been interesting to be able to run unlinked at 497 x 8 to see what the results may have been.

Link to benches here

[mussadek]

Pic #1. 446 x 9 (unlinked) which has lower bandwidth & lower latency is faster in windows

[markost]

Certainly feels that way. Lower Super Pi scores and lower latency gives more speed as against higher bandwidth and higher latency which gives better bandwidth benches.

[ShoNuff]

These straps just keep getting stranger all the time. just threw in some Reaper PC2 9200 and bandwidth is definitely not uniform. When running memtest with memory set at 1t @ 1015Mhz it shows memory bandwidth as 5615 MB/s. Up that to 1025 Mhz and bandwidth jumps to 5815 MB/s. Fair enough one would think. I then raised memory speed to 1035 Mhz, then 1045 Mhz and finally 1056 Mhz all at 1t and 4.4.4.4 with tight subs. Bandwidth stayed at 5815 Mhz on each occasion. Everest Benches confirmed that at all these settings, speed did not change and remained practically identical to the screenshot. Anyone got any theories here ?? 1301 BIOS.

I agree with you 100%! Timings were much faster with the dual core.


Quad 4-4-4-12-1T


Dual 3-3-3-9-1T


Benchmarks are "so-so" but I think things can be improved with a new bios.

[markost]

Hey Shonuff : Are you getting 4.0 Ghz on the quad with watercooling ?. If so, what are your temps.