Guest starkos Posted September 26, 2010 Report Posted September 26, 2010 FYI: I believe your overclocking is just costmetical. I've managed to get noticable benchmark changes now but using a different method, see this thread and links in it: http://android.modaco.com/content/t-mobile...i/#entry1416600 Running my U8100 at 691.2MHz as it did a spontaous reboot in 710.4MHz. The original patch didn't fit our kernel exactly but I think I figured out how to apply the changes to the pulse mini acpuclk_set_div: /* Set proper dividers for the given clock speed. */ static void acpuclk_set_div(const struct clkctl_acpu_speed *hunt_s) { // uint32_t reg_clkctl, reg_clksel, clk_div, src_sel; uint32_t reg_clkctl, reg_clksel, clk_div, src_sel, a11_div; reg_clksel = readl(A11S_CLK_SEL_ADDR); /* AHB_CLK_DIV */ clk_div = (reg_clksel >> 1) & 0x03; /* CLK_SEL_SRC1NO */ src_sel = reg_clksel & 1; a11_div=hunt_s->a11clk_src_div; if(hunt_s->a11clk_khz>=528000) { a11_div=0; writel(0x24, MSM_CLK_CTL_BASE+0x33C); udelay(50); } /* * If the new clock divider is higher than the previous, then * program the divider before switching the clock */ if (hunt_s->ahbclk_div > clk_div) { reg_clksel &= ~(0x3 << 1); reg_clksel |= (hunt_s->ahbclk_div << 1); writel(reg_clksel, A11S_CLK_SEL_ADDR); } /* Program clock source and divider */ reg_clkctl = readl(A11S_CLK_CNTL_ADDR); reg_clkctl &= ~(0xFF << (8 * src_sel)); reg_clkctl |= hunt_s->a11clk_src_sel << (4 + 8 * src_sel); reg_clkctl |= a11_div << (0 + 8 * src_sel); writel(reg_clkctl, A11S_CLK_CNTL_ADDR); /* Program clock source selection */ reg_clksel ^= 1; writel(reg_clksel, A11S_CLK_SEL_ADDR); /* * If the new clock divider is lower than the previous, then * program the divider after switching the clock */ if (hunt_s->ahbclk_div < clk_div) { reg_clksel &= ~(0x3 << 1); reg_clksel |= (hunt_s->ahbclk_div << 1); writel(reg_clksel, A11S_CLK_SEL_ADDR); } } I suppose your code looks pretty much the same so you may be able to use this. What it does is that it hi-jacks the 528MHz setting and changes it. So when it displays 528MHz in setcpu it's really the new speed. Original code hade 0x25 which means 37 in decimal and it gives 37*19.2=710.4MHz but since that didn't work on my phone I decreased it to 36(0x24). Kinda makes sense to me to be honest as I personally haven't noticed a speed up in the phone what so ever but many people did, maybe a placebo effect lol?
Guest Epic-Emodude Posted September 26, 2010 Report Posted September 26, 2010 With overclock I got 10755ms on the long benchmark (dont know how this compares to pre overclock) and I got over 7 hours of stress test out of it, before I had to start using the phone as a Phone
Guest kallt_kaffe Posted September 26, 2010 Report Posted September 26, 2010 With overclock I got 10755ms on the long benchmark (dont know how this compares to pre overclock) and I got over 7 hours of stress test out of it, before I had to start using the phone as a Phone Well I got one long benchmark result under 9000ms @ 691MHz so either way this method seems to be more effective.
Guest starkos Posted September 26, 2010 Report Posted September 26, 2010 Well I got one long benchmark result under 9000ms @ 691MHz so either way this method seems to be more effective. I'm pretty sure your right to be honest! Unfortunately I haven't got a clue what that patch means or how to adapt it ha! If your bored at any time feel free to have a poke around in the Pulse source :lol:
Guest Epic-Emodude Posted September 26, 2010 Report Posted September 26, 2010 Ive just done a bit of fiddleing with it, and it seems that (in terms of long benchmark anyway) 680mhz is optimal speed (can someone back me up on this?) I got 10235 with this speed as opposed to 107something with 780
Guest flip360 Posted September 26, 2010 Report Posted September 26, 2010 FYI: I believe your overclocking is just costmetical. I've managed to get noticable benchmark changes now but using a different method, see this thread and links in it: http://android.modaco.com/content/t-mobile...i/#entry1416600 Running my U8100 at 691.2MHz as it did a spontaous reboot in 710.4MHz. The original patch didn't fit our kernel exactly but I think I figured out how to apply the changes to the pulse mini acpuclk_set_div: /* Set proper dividers for the given clock speed. */ static void acpuclk_set_div(const struct clkctl_acpu_speed *hunt_s) { // uint32_t reg_clkctl, reg_clksel, clk_div, src_sel; uint32_t reg_clkctl, reg_clksel, clk_div, src_sel, a11_div; reg_clksel = readl(A11S_CLK_SEL_ADDR); /* AHB_CLK_DIV */ clk_div = (reg_clksel >> 1) & 0x03; /* CLK_SEL_SRC1NO */ src_sel = reg_clksel & 1; a11_div=hunt_s->a11clk_src_div; if(hunt_s->a11clk_khz>=528000) { a11_div=0; writel(0x24, MSM_CLK_CTL_BASE+0x33C); udelay(50); } /* * If the new clock divider is higher than the previous, then * program the divider before switching the clock */ if (hunt_s->ahbclk_div > clk_div) { reg_clksel &= ~(0x3 << 1); reg_clksel |= (hunt_s->ahbclk_div << 1); writel(reg_clksel, A11S_CLK_SEL_ADDR); } /* Program clock source and divider */ reg_clkctl = readl(A11S_CLK_CNTL_ADDR); reg_clkctl &= ~(0xFF << (8 * src_sel)); reg_clkctl |= hunt_s->a11clk_src_sel << (4 + 8 * src_sel); reg_clkctl |= a11_div << (0 + 8 * src_sel); writel(reg_clkctl, A11S_CLK_CNTL_ADDR); /* Program clock source selection */ reg_clksel ^= 1; writel(reg_clksel, A11S_CLK_SEL_ADDR); /* * If the new clock divider is lower than the previous, then * program the divider after switching the clock */ if (hunt_s->ahbclk_div < clk_div) { reg_clksel &= ~(0x3 << 1); reg_clksel |= (hunt_s->ahbclk_div << 1); writel(reg_clksel, A11S_CLK_SEL_ADDR); } } I suppose your code looks pretty much the same so you may be able to use this. What it does is that it hi-jacks the 528MHz setting and changes it. So when it displays 528MHz in setcpu it's really the new speed. Original code hade 0x25 which means 37 in decimal and it gives 37*19.2=710.4MHz but since that didn't work on my phone I decreased it to 36(0x24). Do i use this code for the patch or should something be changed from the original one at xda?
Guest gusthy Posted September 26, 2010 Report Posted September 26, 2010 (edited) Kinda makes sense to me to be honest as I personally haven't noticed a speed up in the phone what so ever but many people did, maybe a placebo effect lol? Someone wants to look at Linpack results on the web? :lol::) Look at last one B) (I didn't increase to 0x25, maybe someone tries it, but I dont mind it B)) http://www.greenecomputing.com/apps/linpac...pack-by-device/ Edited September 26, 2010 by gusthy
Guest DanWilson Posted September 26, 2010 Report Posted September 26, 2010 Someone wants to look at Linpack results on the web? :lol::) Look at last one B) (I didn't increase to 0x25, maybe someone tries it, but I dont mind it B)) Not the last one, but how the ***k did you get 4.857!?!
Guest starkos Posted September 26, 2010 Report Posted September 26, 2010 Someone wants to look at Linpack results on the web? B):) Look at last one :lol: (I didn't increase to 0x25, maybe someone tries it, but I dont mind it :lol:) LOL B) ! Any chance of providing the patch you actually used as I'm not good with these kind of things & as I use my own rom I would have to make it my self :lol:
Guest gusthy Posted September 26, 2010 Report Posted September 26, 2010 (edited) Not the last one, but how the ***k did you get 4.857!?! I just compiled in the patch recommended above :lol: took 5 minutes. It seems it works. Edit: Say thanks to kallt_kafe Edited September 26, 2010 by gusthy
Guest gusthy Posted September 26, 2010 Report Posted September 26, 2010 (edited) LOL B) ! Any chance of providing the patch you actually used as I'm not good with these kind of things & as I use my own rom I would have to make it my self :lol: Simply replace the function mentioned above in arch/arm/mach-msm/acpuclock.c Edit: currently it boots with 0x25 Edited September 26, 2010 by gusthy
Guest starkos Posted September 26, 2010 Report Posted September 26, 2010 Simply replace the function mentioned above in arch/arm/mach-msm/acpuclock.c 'Replace the function mentioned above' unfortunately means nothing to me :lol: I was hoping it was going to be some kind of .diff patch or something using the code supplied by kallt_kafe lol
Guest gusthy Posted September 26, 2010 Report Posted September 26, 2010 (edited) 'Replace the function mentioned above' unfortunately means nothing to me B) I was hoping it was going to be some kind of .diff patch or something using the code supplied by kallt_kafe lol What kalt did is exactly that: - You open acpuclock.c - Search for acpuclk_set_div - Delete the whole function - Insert the replacement code kalt posted here - Save - Build kernel With 25 Linpack is above 5 :lol:new_boot.img Edited September 26, 2010 by gusthy
Guest starkos Posted September 26, 2010 Report Posted September 26, 2010 What kalt did is exactly that: - You open acpuclock.c - Search for acpuclk_set_div - Delete the whole function - Insert the replacement code kalt posted here - Save - Build kernel With 25 Linpack is above 5 B) I'm so sorry lol, I just realised as you was typing that above is just a modified section of the acpuclock.c :lol: ! Do you think it's possible to modify it so that it adds the new speed as well as keeping 528?
Guest gusthy Posted September 26, 2010 Report Posted September 26, 2010 (edited) I'm so sorry lol, I just realised as you was typing that above is just a modified section of the acpuclock.c :lol: ! Do you think it's possible to modify it so that it adds the new speed as well as keeping 528? I guess yes, but it needs some more coding. Somewhere I've read about it but I don't remember the exact way. Edit: it is faaaaaaaaaaaaast Edit2: 27 is booted too. I felt a little warm, and Opera didn't seem really faster, so I reverted to 25. Just for fun I uploaded an 5.3 LinPack B)) Edited September 26, 2010 by gusthy
Guest Andy Veliz Posted September 26, 2010 Report Posted September 26, 2010 What kalt did is exactly that: - You open acpuclock.c - Search for acpuclk_set_div - Delete the whole function - Insert the replacement code kalt posted here - Save - Build kernel With 25 Linpack is above 5 :lol: did you use flib's git kernel source?
Guest gusthy Posted September 26, 2010 Report Posted September 26, 2010 did you use flib's git kernel source? No, the original Huawei one.
Guest stealthstealer Posted September 26, 2010 Report Posted September 26, 2010 flashable .zip? :lol:
Guest gusthy Posted September 26, 2010 Report Posted September 26, 2010 flashable .zip? :lol: I guess it is all the same to flash this with flashboot. I forgot to mention, it is for Froyo.
Guest rulerofkaos Posted September 26, 2010 Report Posted September 26, 2010 (edited) No, the original Huawei one. I get a bootloop with eclair. edit: ok its for froyo :-) Edited September 26, 2010 by rulerofkaos
Guest gusthy Posted September 26, 2010 Report Posted September 26, 2010 I get a bootloop with eclair. sure, that boot.img is for froyo. you can repack foor eclair.
Guest rulerofkaos Posted September 26, 2010 Report Posted September 26, 2010 sure, that boot.img is for froyo. you can repack foor eclair. Maybe it can be repacked, but I can't :lol:
Guest stealthstealer Posted September 26, 2010 Report Posted September 26, 2010 I guess it is all the same to flash this with flashboot. I forgot to mention, it is for Froyo. you mean fastboot? and yeah i would, but my laptop just refuses to install pulse drivers. :lol: im already on froyo too thanks B)
Guest gusthy Posted September 26, 2010 Report Posted September 26, 2010 you mean fastboot? and yeah i would, but my laptop just refuses to install pulse drivers. :lol: im already on froyo too thanks B) Unfortunately I don't have the tools to make signed zip now on my machine B)
Guest gusthy Posted September 26, 2010 Report Posted September 26, 2010 (edited) Maybe it can be repacked, but I can't :lol: A minute, I will upload one Edit: it is here, with 0x25 multiplier so it is 710 MHz, same as the Froyo version.boot_eclair.img Edited September 26, 2010 by gusthy
Recommended Posts
Please sign in to comment
You will be able to leave a comment after signing in
Sign In Now