Interesting sleep bug with Bloodrage.

I've put PC to sleep, today I started it (wake up) and I was taking a look into Everest and wow.. I have 1400 MHz 4-8-8-24 T1 ram now. :D (in BIOS its set to 8-8-8-24 T1 ofc). Interesting is, that CPU-Z says same, memtest says that everything is fine.. but most interesting is, that I have about 1k more copy in Everest.

Kinda wonder how is this possible and how its possible that it works. :D

Ofc maybe its just reporting wrong numbers and Everest is lying as usually..

Hm, SuperPI is same, I guess its just lying info. :)

Same with me on CPUz atm im running 4 7 7 16 @ 782 MHz. :D

after boot its normal 7 7 7 16 / btw i set 7 7 7 18 in Bios but sometimes it changes to 16 also in Bios.
So BR dont like loose timings

I'm kinda not sure if those are real timings. Best try is to run SuperPI before sleep (with BIOS set timings) and after sleep with those crazy ones. :D

Btw. whats your BIOS version? I'm on P07.

18 changes in 16 after a reboot... everytime...and with all bios version..

With the previous bios versions also some problem with 21 that changes to 20...if i remember...

:shakes:

just to not making one more thread.. i have many of them already :)
i'd like to ask here

g25 bios, turbo always on, eist on, vista profile performance
system running 200x21 1.32v cpu
then i go into sleep s3 and get back
system running 200x20 1.24v cpu :confused:
after restart it is again 200x21 1.32v
this all according readings in cpu-z, everest and aegis on vista fresh install with all latest drivers

does s3 work good with cpu turbo always on?

I'm on bios P07... after sleep, cpu-z reports my cas latency as 4 instead of 8, everything else is normal.

I know my cas is not really at 4, else I'd get a blue screen.

yes, my memory also 1600 7-7-7-20 before sleep and 1600 4-7-7-20 after

but those cpu multi and voltages is smth new

do u have 200x21 before and after sleep?

Its not dangerous bug.. and no effect on performace, just wrong reading.. :) Not sure if it persist to G25 or P08.

by wrong reading you mean bug with memory cas 4 only or with cpu x20 and voltage also?

Quote:

---

Originally Posted by rioja

yes, my memory also 1600 7-7-7-20 before sleep and 1600 4-7-7-20 after

but those cpu multi and voltages is smth new

do u have 200x21 before and after sleep?

---

For me, the only thing reported different after sleep is cas# latency.... performance remains the same.

different cpu multi and vcore after s3 resume?
thats odd... i dont have that with G25... :shrug:

and about cas4, yeah, it really IS cas4!
its just that the extra memory bandwidth is barely used... you all know how everything above 1333 with tight timings doesnt really make a diference...

hehehe :lol:

Quote:

---

Originally Posted by saaya

different cpu multi and vcore after s3 resume?
thats odd... i dont have that with G25... :shrug:

and about cas4, yeah, it really IS cas4!
its just that the extra memory bandwidth is barely used... you all know how everything above 1333 with tight timings doesnt really make a diference...

hehehe :lol:

---

I read an article that anything over 1333mhz will only net 2-3 more fps in gaming

you cant say its 2-3fps since some games are at 50fps and others at 300fps :D
id say it probably adds 1% extra performance...

maybe more if you are pushing the cpu to 5.5G and run tri or quad sli or xfire... but its not much...
tweaking the memory speed and timings is usually the last thing extreme overclockers do when they maxed out all the other hardware, its like the cream topping, possibly giving a few extra points :D but there are quite some people out there who never really max out their mem cause its not worth the hassle for them.

the only exceptions are superpi and everest afaik...

About game vs. memory speed. Try how long it will take for game to load at 1333 CL9 and at 1600 CL8 or 2000 CL7. :) I think you'll be suprised.. faster rams, aren't for more FPS.. but everything works and response faster.. And you have that warm fuzzy feeling, that you have really fast rams .. which I lack cause I have this crappy rams. :D

Quote:

---

Originally Posted by Mescalamba

About game vs. memory speed. Try how long it will take for game to load at 1333 CL9 and at 1600 CL8 or 2000 CL7. :) I think you'll be suprised.. faster rams, aren't for more FPS.. but everything works and response faster.. And you have that warm fuzzy feeling, that you have really fast rams .. which I lack cause I have this crappy rams. :D

---

why would i run 2000 c7 though if i can run 1333 555? :D
and with the micron d9jnl i have i dont even need high vdimm for that :)

and the biggest effect there is probably uncore clocks, not actually memory clocks. try the same uncore clocks with diferent memory clocks and youll see that uncore clocks make up around half the memory performance in synthetic tests :D

Quote:

---

Originally Posted by saaya

why would i run 2000 c7 though if i can run 1333 555? :D
and with the micron d9jnl i have i dont even need high vdimm for that :)

and the biggest effect there is probably uncore clocks, not actually memory clocks. try the same uncore clocks with diferent memory clocks and youll see that uncore clocks make up around half the memory performance in synthetic tests :D

---

Sure, but not sure if CL7 2GHz rams can do CL5 1333. :) Microns can, I know that.. (dunno if KPT can too).

You run D9JNL.. which is only 1GB per stick, right?

And I know that uncore affects that too.. I have it boosted up, but even higher uncore can't lower latency.. :) Only high speed rams can.

yepp, only 3gb... but d9jnl 2gb sticks are available and cheap afaik,. and they are limited in speed and vdimm they accept and still scale, but not timings.
ive had 2gb d9jnl sticks at 1333 555 running fine iirc... or was it even 544?

so elpida and samsung cant run cas5?
that sucks...

elpida? cas5? walk in the park :D

http://img268.imageshack.us/img268/1...9072717193.png

i can't get my g-skill pi(samsung chip) to run 1333 cas 5 tho, and volts doesn't help

I have not noticed timing changes BUT...

I have read about and confirmed S3 sleep bug where if you go to S3 sleep, resume, then re-enter S3 sleep, unit will not wake up. After hard power down, and unplug, attempts to re-start fail with Blue LED, Force RESET, CMOS clear, PWR/RST flashing. (BIOS G25\P06)

Unfortunately system went to sleep today twice as I hadn't changed appropriate settings until after this mess occurred.

Any idea what this error mode is? I performed CMOS clear with jumper, but still no boot until one of the times I switched PS off and the Force RESET light went steady so I powered on PS immediately (before Caps discharged) and powered on unit OK. Subsequent power off of PS resulted in same state, now only restarts after shutdown if PS switch left on. (so far)

Bad PS or on board power? Does my CMOS Battery need recharge time?

Any helpful comments appreciated.

1600 cas5 mmhhhhhh, yummy! :D

mr fixit, i went to s3 and resumed several times without problems...
im not running oced though atm... :shrug:
im on G25 and vista32...

You are brave to try....Love your avatar BTW.

Ever seen blinking lights error state?

Quote:

---

Originally Posted by MrFixit

You are brave to try....Love your avatar BTW.

Ever seen blinking lights error state?

---

yeah, if i turn off the psu for a couple of hours and then turn it on again and start the pc it tends to get stuck at some point, so i have to hit the rest button and then it boots fine...

but besides that all is good...

Quote:

---

Originally Posted by MrFixit

I have not noticed timing changes BUT...

I have read about and confirmed S3 sleep bug where if you go to S3 sleep, resume, then re-enter S3 sleep, unit will not wake up. After hard power down, and unplug, attempts to re-start fail with Blue LED, Force RESET, CMOS clear, PWR/RST flashing. (BIOS G25\P06)

Unfortunately system went to sleep today twice as I hadn't changed appropriate settings until after this mess occurred.

Any idea what this error mode is? I performed CMOS clear with jumper, but still no boot until one of the times I switched PS off and the Force RESET light went steady so I powered on PS immediately (before Caps discharged) and powered on unit OK. Subsequent power off of PS resulted in same state, now only restarts after shutdown if PS switch left on. (so far)

Bad PS or on board power? Does my CMOS Battery need recharge time?

Any helpful comments appreciated.

---

Not sure if Im correct, but this can by maybe CPU issue, and its described somewhere in i7 errata.. and you can probably even RMA that CPU.. tough it should be only pre-D0 stepping issue..?

i also get the cas4 bug lol
but it always wakeup properly tho even after multiple sleep/wakeup
http://img528.imageshack.us/img528/2...9072809052.png

I have the same issue, keeps me from sleeping the machine.
I have a G15 keyboard, and on resume (when it won't resume) it causes the lights to flicker very quickly, looks to be at about 60 Hz.

Quote:

---

Originally Posted by MrFixit

I have not noticed timing changes BUT...

I have read about and confirmed S3 sleep bug where if you go to S3 sleep, resume, then re-enter S3 sleep, unit will not wake up. After hard power down, and unplug, attempts to re-start fail with Blue LED, Force RESET, CMOS clear, PWR/RST flashing. (BIOS G25\P06)

Unfortunately system went to sleep today twice as I hadn't changed appropriate settings until after this mess occurred.

Any idea what this error mode is? I performed CMOS clear with jumper, but still no boot until one of the times I switched PS off and the Force RESET light went steady so I powered on PS immediately (before Caps discharged) and powered on unit OK. Subsequent power off of PS resulted in same state, now only restarts after shutdown if PS switch left on. (so far)

Bad PS or on board power? Does my CMOS Battery need recharge time?

Any helpful comments appreciated.

---