Page 5 of 6 FirstFirst ... 23456 LastLast
Results 101 to 125 of 133

Thread: Howto disable the tlb-fix with CrystalCPU

  1. #101
    Xtreme Enthusiast Raiderman's Avatar
    Join Date
    Feb 2008
    Location
    Idaho
    Posts
    503
    Quote Originally Posted by LIKMARK View Post
    Yes, in the same directory and tried running them as admin, no luck though. Vista x64.
    I had the same error while trying to execute the file while it was still packed in the rar file. If you extract it to the same dir as crystal cpu, then right click, run as admin, everything should work perfectly.

  2. #102
    Xtreme Member
    Join Date
    Nov 2007
    Location
    philippines
    Posts
    279
    Quote Originally Posted by Raiderman View Post
    I had the same error while trying to execute the file while it was still packed in the rar file.
    you were to run while it was still packed?!
    AMD Athlon 64 x2 6000+ AM2 CCB8F 0740 FPMW
    MSI K9A2 Platinum v1.2
    4 x 1GB Corsair XMS2 (rev 5.1 and rev 5.2)
    Zalman CNPS9700LED
    PowerColor ATi HD 3870 512MB DDR4 256bit PCIe 2.0
    Corsair TX750 Watts (12v @ 60A on a single rail)
    80 GB WD SATA I (primary)
    250 GB WD SATA II (backup)




  3. #103
    Xtreme Enthusiast Raiderman's Avatar
    Join Date
    Feb 2008
    Location
    Idaho
    Posts
    503
    LOL I know I was in a hurry
    C3 Phenom II 965 BE (0941CPMW)
    Msi 790Fx-GD 70
    2x2 gb OCZ AMD Black Edition 1600mhz DDR 3
    2x MSI HD 5850's OC edition X-fire
    2x Western Digital 250gb 16mb cache Raid-0
    1x Seagate 1 TB Data
    SILVERSTONE|DA700 700W RT Power Supply
    Koolance INX-720 (Koolance 345 block)
    LianLi pc-60fwb




  4. #104
    Xtreme Enthusiast Raiderman's Avatar
    Join Date
    Feb 2008
    Location
    Idaho
    Posts
    503
    Is there a way to disable the dialogue box from popping up on startup. It would be cool to have the command window open at startup, apply the fix and then disapear.

    @sam, thanks for your hard work, its much appreciated!
    @KTE. thanks also
    C3 Phenom II 965 BE (0941CPMW)
    Msi 790Fx-GD 70
    2x2 gb OCZ AMD Black Edition 1600mhz DDR 3
    2x MSI HD 5850's OC edition X-fire
    2x Western Digital 250gb 16mb cache Raid-0
    1x Seagate 1 TB Data
    SILVERSTONE|DA700 700W RT Power Supply
    Koolance INX-720 (Koolance 345 block)
    LianLi pc-60fwb




  5. #105
    Xtreme Member
    Join Date
    Nov 2007
    Location
    philippines
    Posts
    279
    Quote Originally Posted by Raiderman View Post
    LOL I know I was in a hurry
    lolz! excited you mean...how was your OC?
    AMD Athlon 64 x2 6000+ AM2 CCB8F 0740 FPMW
    MSI K9A2 Platinum v1.2
    4 x 1GB Corsair XMS2 (rev 5.1 and rev 5.2)
    Zalman CNPS9700LED
    PowerColor ATi HD 3870 512MB DDR4 256bit PCIe 2.0
    Corsair TX750 Watts (12v @ 60A on a single rail)
    80 GB WD SATA I (primary)
    250 GB WD SATA II (backup)




  6. #106
    Xtreme Enthusiast
    Join Date
    Feb 2007
    Location
    Norway
    Posts
    704
    Quote Originally Posted by sam2008 View Post
    Is your error message "Cannot load DLL!". Can you find the Sysinfo.DLL in the same directory? I myself use x64, so should be no problem.
    That's the message yes. You mean the sysinfox64.dll I presume? Yes I can find it, but I can't find a sysinfo.dll. No

  7. #107
    Registered User
    Join Date
    Feb 2008
    Posts
    47
    Quote Originally Posted by LIKMARK View Post
    That's the message yes. You mean the sysinfox64.dll I presume? Yes I can find it, but I can't find a sysinfo.dll. No
    No, I mean sysinfo.dll. Try redownload the CrystalCPUID, make sure sysinfo.dll is in there. Then you should be fine.
    Last edited by sam2008; 03-06-2008 at 10:17 AM.

  8. #108
    Xtreme Enthusiast
    Join Date
    Feb 2007
    Location
    Norway
    Posts
    704
    ok, had to download the x86 version. not the x64 version. Then it worked. Gonna restart now and see
    Tested winrar right away and got 1200kb/s against 240 with tlb patch enabled. Thanks

  9. #109
    Xtreme Mentor
    Join Date
    May 2007
    Posts
    2,799
    sam2008: Is there any way to control which cores the MSR is applied to in your app through the ini?

    It's because I (and many) oc at bootup settings where 1 core will fail but the rest of 3 won't, to then change the MSR for that weak core to drop multiplier one notch and let it run at lower MHz while the other 3 run at bootup MHz. This isn't the hardest bit, but the bit I need is, when CnQ is enabled I used your app to drop to 0.850VID 800MHz on all cores, but with load, they all need to get back up to the multi's I set initially and not all the same multi (which is what they do unless I change MSR for each core separately).

    So say I use 13x on first 3 cores and 12.5x multi on the last core, even with CnQ enabled, I'd want it to remain max 12.5x under load (or it'd crash).

    Is there any way yet? This'll come in handy for you, test your Phenom, it will very likely have one core very weak compared to the rest.

  10. #110
    Registered User
    Join Date
    Feb 2008
    Posts
    47
    KTE: You word is my command. One way to do it is to add a core mask option in the ini file. I.E. 0b1011 stands for enabling writing to core 1, 2, and 4, but not core 3. How does that sound to you?

  11. #111
    Xtreme Mentor
    Join Date
    May 2007
    Posts
    2,799
    Sounds perfect, eager to test.
    We use "-coremask=x" to disable a core using CPU-Z to avoid the weak core and we also tried disabling a specific core using Bus:0h Dev:24h Fct:3h Offset:BEh 0000000000000000b [core 3,2,1,0], both of them never worked as we wanted so I suppose it is similar to both methods in a way but to get something better done. A DIY replacement for BIOS multi.

  12. #112
    Registered User
    Join Date
    Feb 2008
    Posts
    47
    KTE: I have uploaded the ver1.02 at POST #90. Try to keep all the updates in one post so that it won't be scatter all around. You can download from there. Basically added support for core masking in the INI file. Let me know how it worked out for you .

    Now you can use the COREMASK to specify which core you want the MSR value to be written. It is 4-bit binary value. The lowest bit is core 1, second bit is core 2, third bit is core 3, and fourth bit is core 4. When a bit is set to 1, MSR for that core will be written. If a bit is set to zero, the program will skip the MSR for that core. For example, 0b1011 means core 1, 2, and 4 will be written, and core 3 will be skipped. If COREMASK is set to 0b1111, then the selected MSR for all four cores will be written. If COREMASK is set to 0b0000, then selected MSR won't be written to any of the cores. Changing the COREMASK value accordingly. It is set at 0b1111 by default (all 4 core enabled).

  13. #113
    Registered User
    Join Date
    Feb 2008
    Posts
    47
    I usually like something to indicate that all things are working fine. Otherwise, the console window will just flash and disappear not really know what's going on. That's why I deliberately added the message box. But as you wish, I updated the program, now you can choose to disable the popup message box by adding the -nopop argument at the end of the program. If no argument is provided, the box will popup by default. Go to RebelsHeavenForum to download the TLB ver1.03. In Vista, you have to add the argument -nopop thru the taskscheduler. Hope it works for you.

    Quote Originally Posted by Raiderman View Post
    Is there a way to disable the dialogue box from popping up on startup. It would be cool to have the command window open at startup, apply the fix and then disapear.

    @sam, thanks for your hard work, its much appreciated!
    @KTE. thanks also

  14. #114
    Xtreme Mentor
    Join Date
    May 2007
    Posts
    2,799
    Sam2008: thank you!

    Works perfect. However quick qs: does coremask value override the mask_high/mask_low parameter?

    What I mean is if I set:

    index=0xc0010070
    high=0x00000000
    low=0x5000340A
    mask_low=0xFFFFFFFF
    mask_high=0xFFFFFFFF
    coremask=0b0111

    Will it write 0x5000340A to core4 or skip it reading the mask_low=0xF?

    Could you not use mask_low/mask_high values 0/1/2/3/4/5/6/7/8/9/A/B/C/D/E/F to do the same?
    i.e. using the above, me wanting to write to core 3 (340A) only=>

    index=0xc0010070
    high=0x00000000
    low=0x5000340A
    mask_low=0xFFFFF777 (mask 0111b)
    mask_high=0xFFFFFFFF

    Would this work?

    TBH only now has Phenom really become insanely tweakable and fun to play with.

  15. #115
    Xtreme Member
    Join Date
    Oct 2004
    Posts
    171
    I updated my RW-Everything batch file and now it's using AND to clear only the needed bits and not changing the rest of the msr register.
    I used the new version 0.29 of RW-Everything.

    Code:
    >cpu 1
    read register to LocalD:LocalA
    >rdmsr 0xc0010015
    set bit#4 to zero
    >LocalA = AND LocalA 0xfffffff7
    write back
    >wrmsr 0xc0010015 LocalD LocalA
    read register to LocalD:LocalA
    >rdmsr 0xc0011023
    set bit#2 to zero
    >LocalA = AND LocalA 0xfffffffd
    write back
    >wrmsr 0xc0011023 LocalD LocalA
    
    >cpu 2
    read register to LocalD:LocalA
    >rdmsr 0xc0010015
    set bit#4 to zero
    >LocalA = AND LocalA 0xfffffff7
    write back
    >wrmsr 0xc0010015 LocalD LocalA
    read register to LocalD:LocalA
    >rdmsr 0xc0011023
    set bit#2 to zero
    >LocalA = AND LocalA 0xfffffffd
    write back
    >wrmsr 0xc0011023 LocalD LocalA
    
    >cpu 3
    read register to LocalD:LocalA
    >rdmsr 0xc0010015
    set bit#4 to zero
    >LocalA = AND LocalA 0xfffffff7
    write back
    >wrmsr 0xc0010015 LocalD LocalA
    read register to LocalD:LocalA
    >rdmsr 0xc0011023
    set bit#2 to zero
    >LocalA = AND LocalA 0xfffffffd
    write back
    >wrmsr 0xc0011023 LocalD LocalA
    
    >cpu 4
    read register to LocalD:LocalA
    >rdmsr 0xc0010015
    set bit#4 to zero
    >LocalA = AND LocalA 0xfffffff7
    write back
    >wrmsr 0xc0010015 LocalD LocalA
    read register to LocalD:LocalA
    >rdmsr 0xc0011023
    set bit#2 to zero
    >LocalA = AND LocalA 0xfffffffd
    write back
    >wrmsr 0xc0011023 LocalD LocalA
    
    >rwexit

  16. #116
    Registered User
    Join Date
    Feb 2008
    Posts
    47
    Coremask determines which core's MSR you want to write. Mask_low, and Mask_high determines which bit in the MSR you want to write. Coremask has a higher priority. If you set coremaks=binary 0111, then it will skip any writing operation to core 4.

    If you set
    index=0xc0010070
    high=0x00000000
    low=0x5000340A
    mask_low=0xFFFFFFFF
    mask_high=0xFFFFFFFF
    coremask=0b0111

    You will skip core 4. Nothing will done to core 4. Also since you set the mask_low and mask_high=0xffffffff, so you mask all bits in that MSRs in the other 3 cores. So no change will be made to the MSRs in core 1,2,3 (or actually, it will read the original values in the MSRs in core 1,2,3, and write the original value back in). If you set mask_low and mask_high=0x00000000, then MSRs in core 1,2,3 will be written with 0x5000340A. Still nothing will be done to core 4 because of coremask=0111(b) told the program to skip it.

    The mask_high and mask_low are only useful if you want certain bits in the MSR not to be change. If you want to directly write the specified value into the MSRs, you should set them to 0x00000000.

    For example, if the original low value of a MSR is 0x10000203, and you want change the its 1-4 bits to 9, and 60-64 bits to 4, but do not want to touch the bits 5-59 and want to keep them the way they are originally. Then you can use the mask_low=0x0FFFFFF0, and write the low=0x4BBBBBB9 (B represents a non-care value since it will be masked off) into that MSR. The final value of MSR after the operation will be 0x40000209.

    If you just want to write a specified value directly into MSRs, then set the Mask_high, mask_low=0x0.

    Quote Originally Posted by KTE View Post
    Sam2008: thank you!

    Works perfect. However quick qs: does coremask value override the mask_high/mask_low parameter?

    What I mean is if I set:

    index=0xc0010070
    high=0x00000000
    low=0x5000340A
    mask_low=0xFFFFFFFF
    mask_high=0xFFFFFFFF
    coremask=0b0111

    Will it write 0x5000340A to core4 or skip it reading the mask_low=0xF?

    Could you not use mask_low/mask_high values 0/1/2/3/4/5/6/7/8/9/A/B/C/D/E/F to do the same?
    i.e. using the above, me wanting to write to core 3 (340A) only=>

    index=0xc0010070
    high=0x00000000
    low=0x5000340A
    mask_low=0xFFFFF777 (mask 0111b)
    mask_high=0xFFFFFFFF

    Would this work?

    TBH only now has Phenom really become insanely tweakable and fun to play with.

  17. #117
    Registered User
    Join Date
    Feb 2008
    Posts
    47
    Yeh, RW-Everything is good tool. But it uses a unsigned driver. So in Vista, it will cause my Vista go into a so called "TEST MODE" (driver signing forced to disable) with watermarks over the desktop.

    Quote Originally Posted by mibo View Post
    I updated my RW-Everything batch file and now it's using AND to clear only the needed bits and not changing the rest of the msr register.
    I used the new version 0.29 of RW-Everything.

  18. #118
    Xtreme Enthusiast Raiderman's Avatar
    Join Date
    Feb 2008
    Location
    Idaho
    Posts
    503
    Quote Originally Posted by sam2008 View Post
    I usually like something to indicate that all things are working fine. Otherwise, the console window will just flash and disappear not really know what's going on. That's why I deliberately added the message box. But as you wish, I updated the program, now you can choose to disable the popup message box by adding the -nopop argument at the end of the program. If no argument is provided, the box will popup by default. Go to RebelsHeavenForum to download the TLB ver1.03. In Vista, you have to add the argument -nopop thru the taskscheduler. Hope it works for you.

    WOW! You Rock! Usually I only get it my way at Burger King
    C3 Phenom II 965 BE (0941CPMW)
    Msi 790Fx-GD 70
    2x2 gb OCZ AMD Black Edition 1600mhz DDR 3
    2x MSI HD 5850's OC edition X-fire
    2x Western Digital 250gb 16mb cache Raid-0
    1x Seagate 1 TB Data
    SILVERSTONE|DA700 700W RT Power Supply
    Koolance INX-720 (Koolance 345 block)
    LianLi pc-60fwb




  19. #119
    Xtreme Mentor
    Join Date
    May 2007
    Posts
    2,799
    Yeah I understand that much and have been doing it but was thinking if Fh = 1111b, so you chose to mask the bit, but 0h = 0000b so you chose to not mask those, then maybe (I was hoping) we could use.. never mind

    It won't work, since bits don't stand for each core.

  20. #120
    Registered User
    Join Date
    Feb 2008
    Posts
    47
    @KTE, I know it is a bit confusing, using 1 stand for disabling the MSR bit, but also use 1 for enabling cores. That's all because MSR masking was originally based on the convention set by justapost .

    So does it work out the way you wanted now? Crank up the multiplier for 1,2 and 4 and lower the multi for defective core 3? Just wondering can't you adjust that in AOD? I guess AOD probably won't let you save the settings.

    Quote Originally Posted by justapost View Post
    Hmm tht ini file approach looks nice and easy. Best would be if you can define a mask and a value for a variable number of MSR registers.

    Example:
    Code:

    [MSR]
    C0010015_value=0x00000008
    C0010015_mask=0x00000008
    [MSR0]
    C0010064_value=0x2C0020A0
    C0010064_mask=0x00ff0000

    All stuff in MSR would be applied to all cores stuff in MSR0-MSR3 whould be core specific.
    App must read the MSR register, apply the mask and add add the value.
    Last edited by sam2008; 03-07-2008 at 07:11 AM.

  21. #121
    Xtreme Addict
    Join Date
    Sep 2007
    Location
    Munich, DE
    Posts
    1,401
    I understand the question like that:

    If you apply an value to an MSR on multiple cores, do you perform the masking for each core or do you use the result from the first core for the other cores?

    BTW I like the result a lot better than my proposal. I guess coremask might also work on multi socket machines.

  22. #122
    Registered User
    Join Date
    Feb 2008
    Posts
    47
    Masking is performed separately for each core's MSR's original value since I realized in certain circumstances the bit values for MSR for different cores might be different.

    Quote Originally Posted by justapost View Post
    I understand the question like that:
    If you apply an value to an MSR on multiple cores, do you perform the masking for each core or do you use the result from the first core for the other cores?
    Last edited by sam2008; 03-07-2008 at 12:54 PM.

  23. #123
    Registered User
    Join Date
    Feb 2008
    Posts
    47
    @KTE, any exciting results on separately adjusted multipliers for each core ?

  24. #124
    Xtreme Mentor
    Join Date
    May 2007
    Posts
    2,799
    Sorry, Phenom=no boot for many hours yet. It happens randomly when I reset BIOS on my Phenom as it did on my 1st one.

    So I've not checked. I know 2700x3/2600x1 is fully stable though, not tried more yet. 2665 will also be stable on core3 and 2754 is max I know is stable with core 0,1,2 so I'll give it a try for more 100% stability. Around 222x12.5 3 cores and 222x12 on last core next stop.

  25. #125
    Registered User
    Join Date
    Sep 2007
    Posts
    1
    Thank U sam2008, great job!

Page 5 of 6 FirstFirst ... 23456 LastLast

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •