I suggest you to create a batch file to run CBId with parameters. Sets clock frequency of the DRAM controller and HyperTransport link to 233MHz and 500MHz respectively. Sets tRC and tRFC timings to 11T and 14T respectively
#IPICTURE 1.2 DOWNLOAD HOW TO#
Here is some explanation how to use them with CBId. TCAS, tRCD, tRP, tRAS, tRC, tRFC, tRTP, tRRD, tWTR, tWR, t2T, DCTFID, HTLFID
#IPICTURE 1.2 DOWNLOAD FULL#
Below is the full list of supported command line parameters: Now it is possible to use CBId as a tool to change DRAM timings and some frequencies at startup. I have just finished a new CBId Black Edition beta 3 (). There are a lot of articles about PR Ratings at X-bit labs here (), the first one gives what is supposed to be the new formula for calculating Athlon64 PR Ratings. So by rights it should be a 2900+, not a 3000+! :rolleyes:īTW, the general PR formula for a 133 fsb tbred b is Rating = 3 x Frequency/2 – 500. That runs at 13x166 which is the same as the 2700+.
The trouble is, AMD screwed with the ratings when the 3000+ came out. Exactly the same fsb and MHz, but the extra cache is worth 200 PR points. So (according to AMD) the difference between a 133 and 166 fsb is 50MHz of cpu speed, which is about 75 PR points.Īlso compare the 2600+ (12.5x166) and the 2800+ Barton (12.5x166). when the tbred b's first came out there were 2 variants of the 2600+, 12.5x166.66. Both higher fsb and larger cache size will give a higher PR rating for the same MHz.Į.g. Isn't it possible to calculate PR-rating from frequencies alone?No. Thanks for screenshot! Heh, I would really like to drink a beer with you ) If you're ever in London PM me and we'll see what can be arranged. You can use Escape key for closing the window.ESC key is fine, but I told you it wasn't obvious! :D How about a "ESC to quit" label then? :D Perhaps it was just a marketing feature, I can't see why you would ever want it Enabled on a desktop system anyway (although that was in fact the default setting).Īs for your suggestions about close button I do not plan to add it at all, because it looks inappropriate. I just know that it was implemented in the Abit KX7-333 bios, and Disabled did make a big difference in all benchmarks. I analyzed all Athlon XP MSRs, but I didn’t find how “L2 ECC Cache Checking†BIOS option affect to the set of MSRs.No idea either. Also, I planed to realize the same functions for L2 Cache. Perhaps, it will be a good thing for experiments. I found that all AMD Athlon Model 1/2 has enabled ECC, but in the Athlon XP the correction was disabled. Thanks for screenshot! Heh, I would really like to drink a beer with you )įrankly speaking, I don’t know how L1 Cache ECC affects to the processor performance. Thank you MrLavender, I have just fixed this bug! Hm, you are right about the XP rating! The program incorrectly calculated it for Barton with 200MHz bus. So, my formula isn’t so exact as you can see.
For example, if your CPU has non-typical 170MHz FSB XP rating is calculated like for 166MHz, if 220MHz – like for 200MHz. It’s calculated depending on L2 Cache, standard FSB, CPU clock. You can use Escape key for closing the window. I analyzed all Athlon XP MSRs, but I didn’t find how “L2 ECC Cache Checking†BIOS option affect to the set of MSRs.Īs for your suggestions about close button I do not plan to add it at all, because it looks inappropriate. Here is my 2500+ at 3200+ speed (very slightly over).įrankly speaking, I don’t know how L1 Cache ECC affects to the processor performance. The XP rating is calculated a bit too high as somebody else pointed out. Or if there is one there already that I've missed then make it more obvious!Ģ. If you are the author, here are my suggestions ġ. Secondly, my program shows XP-Rating for all Athlons XP and your CPU isn’t an exception to the rule.You are the author of this program? Nice util, but I wouldn't suggest that accusing ppl of posting fake pics is a good way to collect bug reports.:( Your screenshot is fake! At fist, 128KB L2 cache has your Athlon XP. I specifically put L2 in bold in my reply to highlight that difference. The inference being that you want L1 cache checking disabled for best performance as well. Somebody asked what would the effect of enabling/disabling L1 cache checking, I pointed out that disabling L2 checking used to give a performance boost on mobos that allowed that. Be attentive! Erm, I'm very attentive thank you. The program can disable or enable L1 Cache ECC only.
0 kommentar(er)
