CBB有史以來的第一件以錄影方式介紹硬體及分享(Part I)
- 主題發起人 狂少
- 開始日期
- 已加入
- 9/18/03
- 訊息
- 14,397
- 互動分數
- 2
- 點數
- 0
後...我真的是老了
撐一天沒睡,再睡一天@@
說起來,這是一種即興混亂演示...
下次我要改善成好一點的,
等我搬家弄好一點..我就能抓一個角度然後架3角架
第3段 和第5段剛剛給gas1123兄...感謝他的努力分流
3,5有提到我和日本人對Super pi方法之不同應用
真的還算蠻清楚
日本人真是很小器
表面上是有說明,但是其中道理卻是隻字未提...
最重要是為達Copy waza的效果...還笨笨的繞一大圈
更笨的是XS那群哈日族--Vapor, Gautum還一直說那是聖經...真是笑到我肚子痛
撐一天沒睡,再睡一天@@
說起來,這是一種即興混亂演示...
下次我要改善成好一點的,
等我搬家弄好一點..我就能抓一個角度然後架3角架
第3段 和第5段剛剛給gas1123兄...感謝他的努力分流
3,5有提到我和日本人對Super pi方法之不同應用
真的還算蠻清楚
日本人真是很小器
表面上是有說明,但是其中道理卻是隻字未提...
最重要是為達Copy waza的效果...還笨笨的繞一大圈
更笨的是XS那群哈日族--Vapor, Gautum還一直說那是聖經...真是笑到我肚子痛
- 已加入
- 9/18/03
- 訊息
- 14,397
- 互動分數
- 2
- 點數
- 0
這底下是我之前在OC team那邊貼的...
如果大家看了地3段和第5段就會一目了然
但是這只是日本人的copy waza改良,尚不是狂少所用來算super pi的密訣
Dear all
this tweak I would like to share today, it probably is gonna be the real weapon of OPB Super pi tweaks.
First of , we all know about some key tweaks about pi, such as registry tweakings, paging files size,also
the last one,
the Copy waza tweaking.
Actually, personally, copy waza is really based on the very great concept of the fact by adding a virtual
cache for system which we are
carefully design for running super pi especially very effective for restricting the size of system memory
usage to 600mb~800mb intentionally.
ok perhaps some of you understand what is all about, but here you might not see through that.
Certainly , here the System memory usage I am talking about is equal (virtually) to system's physical
memory, note, here the system physical memory WILL BE equal to the available memory plus system cache ;
moreover, this condition MUST be only and ONLY if you manually set the maxmem to the range we talk about
earlier.. For this theory, you can check it from the task manager, please see the screen captures below.
so, after understand the outline of those key elements, , now, we need to understand more aboout how to set
up copy waza work for OPB cache diversing tweak
Here I would like to sayis, this theory is not restricted you must set the exactly numbers as I do here,
however, after long term tried and experiment; all these number and the registry tweak settings (please
click here) match -up are the best settings integrated for me.
As my tweaking sequence will be as followed:
1.Clean install, i.e. delete OS partition and re-create the partition again then use regular slow format
during windows installation.
1a. integrate the cleaner (clean the temp install files after OS finish installation) and updated files from
MS.
1b. Integrate the driver you need
2.install all drivers if you don't do 1b.
2a. apply the reg file I post.
3. set pagaing file from 512mb to 512mb place at the partition other than your OS partition
add maxmem=600 or 624 (616 is also ok)... then reboot.
4. confirm you physical memory by checking that from task manager just show @ above I post, not exactly need
to be the same. but once if you set maxmem parameters, yuour available + system cache always about maching
your virtually total of system ram (maxmem)
5. prepare the copy waza file ,
make a single file (rar) about 1/2 of your real installed memory , should be about double of your paging
file. place it on any partition then move to your OS partition.
reboot.
6. run super pi after system boot in and idle for about 15~30 seconds. then run twice of 16k to clear super
pi cache. then it's up to you wanna run 1m once orr not. now don't close the super pi window and get it on
the status before you press the OK to run 32m.(from tab calculate---> select digits of pi to be calculate--
->32m--->now start .... but don't press the ok)
run task manager by right click the task bar, not pressing the ctrl+del+alt..., got to performance tab, then
do the copy waza and check the size of system available and system cache memory size changing status. you
should see available memory will be immediately decrease, on the other hand, system cache will boost close
to your maxmem.. but after everytime you done for CW, the system memory will be back from the initial normal
size and even more than that. At that moment, that's the time for start super pi 32m. this is very very
important, you must wait until the sum of system available+ cache memory way greater than system physical
memory (maxmem)
OK, until now, you might think these are NOT what your tweak!! oh ya, it's not. but if you make wrong step
or wrong settings , then you can not get the double accleration (maybe not enough to call double), below, I
will explain why my very ending part of super pi always get great boost than others.
that's right , in order to run spi faster than normal, Japanese team use maxmem and copy waza to get system
cache size biger and biger. This idea is from the windows server 2003 system properties advance setting that
let system use more memory at system cache. Now, as we know, after you obtain that much cache. where does it
go? and how?
I can't say only I do this, but, as I have spending so long time to study this, super pi calculation is
divided by 3 parts or group. First group I name it as Group A, from loop 1~10 or even 11, the second parts
or group B is from 11~20 (for sure it's 20, can't be 21, you will know later), then the last part is from
21~the end, but also can be 20~23. because 23~24 decide to know what the calculation time 24~finish loop
will take. and more importantly, please remember one thing so you can understand the whole topic.
"Super pi process can be going into 3 states as the sequence will be:
Regular calculation speed----> variable calculation speed---->Acclerating calcuation Speed"
so this is how my tweaks developed; you can't change the first and second state a lot. however, trust me, if
you do things right, the third part is gonna be the fatal key to win over peeps
Consequently, , here, I will tell you how OPB Diversing Cache will do for you to get the acclerating part do
the job. Let's see the explain screeny first .
this is the whole tweak except two things
one is copy waza and the last one, let's call OPB-CDT for spi.
then what would happened if copy waza aplly correctly? the result will be about the same as this:
如果大家看了地3段和第5段就會一目了然
但是這只是日本人的copy waza改良,尚不是狂少所用來算super pi的密訣
Dear all
this tweak I would like to share today, it probably is gonna be the real weapon of OPB Super pi tweaks.
First of , we all know about some key tweaks about pi, such as registry tweakings, paging files size,also
the last one,
the Copy waza tweaking.
Actually, personally, copy waza is really based on the very great concept of the fact by adding a virtual
cache for system which we are
carefully design for running super pi especially very effective for restricting the size of system memory
usage to 600mb~800mb intentionally.
ok perhaps some of you understand what is all about, but here you might not see through that.
Certainly , here the System memory usage I am talking about is equal (virtually) to system's physical
memory, note, here the system physical memory WILL BE equal to the available memory plus system cache ;
moreover, this condition MUST be only and ONLY if you manually set the maxmem to the range we talk about
earlier.. For this theory, you can check it from the task manager, please see the screen captures below.
so, after understand the outline of those key elements, , now, we need to understand more aboout how to set
up copy waza work for OPB cache diversing tweak
Here I would like to sayis, this theory is not restricted you must set the exactly numbers as I do here,
however, after long term tried and experiment; all these number and the registry tweak settings (please
click here) match -up are the best settings integrated for me.
As my tweaking sequence will be as followed:
1.Clean install, i.e. delete OS partition and re-create the partition again then use regular slow format
during windows installation.
1a. integrate the cleaner (clean the temp install files after OS finish installation) and updated files from
MS.
1b. Integrate the driver you need
2.install all drivers if you don't do 1b.
2a. apply the reg file I post.
3. set pagaing file from 512mb to 512mb place at the partition other than your OS partition
add maxmem=600 or 624 (616 is also ok)... then reboot.
4. confirm you physical memory by checking that from task manager just show @ above I post, not exactly need
to be the same. but once if you set maxmem parameters, yuour available + system cache always about maching
your virtually total of system ram (maxmem)
5. prepare the copy waza file ,
make a single file (rar) about 1/2 of your real installed memory , should be about double of your paging
file. place it on any partition then move to your OS partition.
reboot.
6. run super pi after system boot in and idle for about 15~30 seconds. then run twice of 16k to clear super
pi cache. then it's up to you wanna run 1m once orr not. now don't close the super pi window and get it on
the status before you press the OK to run 32m.(from tab calculate---> select digits of pi to be calculate--
->32m--->now start .... but don't press the ok)
run task manager by right click the task bar, not pressing the ctrl+del+alt..., got to performance tab, then
do the copy waza and check the size of system available and system cache memory size changing status. you
should see available memory will be immediately decrease, on the other hand, system cache will boost close
to your maxmem.. but after everytime you done for CW, the system memory will be back from the initial normal
size and even more than that. At that moment, that's the time for start super pi 32m. this is very very
important, you must wait until the sum of system available+ cache memory way greater than system physical
memory (maxmem)
OK, until now, you might think these are NOT what your tweak!! oh ya, it's not. but if you make wrong step
or wrong settings , then you can not get the double accleration (maybe not enough to call double), below, I
will explain why my very ending part of super pi always get great boost than others.
that's right , in order to run spi faster than normal, Japanese team use maxmem and copy waza to get system
cache size biger and biger. This idea is from the windows server 2003 system properties advance setting that
let system use more memory at system cache. Now, as we know, after you obtain that much cache. where does it
go? and how?
I can't say only I do this, but, as I have spending so long time to study this, super pi calculation is
divided by 3 parts or group. First group I name it as Group A, from loop 1~10 or even 11, the second parts
or group B is from 11~20 (for sure it's 20, can't be 21, you will know later), then the last part is from
21~the end, but also can be 20~23. because 23~24 decide to know what the calculation time 24~finish loop
will take. and more importantly, please remember one thing so you can understand the whole topic.
"Super pi process can be going into 3 states as the sequence will be:
Regular calculation speed----> variable calculation speed---->Acclerating calcuation Speed"
so this is how my tweaks developed; you can't change the first and second state a lot. however, trust me, if
you do things right, the third part is gonna be the fatal key to win over peeps
Consequently, , here, I will tell you how OPB Diversing Cache will do for you to get the acclerating part do
the job. Let's see the explain screeny first .
this is the whole tweak except two things
one is copy waza and the last one, let's call OPB-CDT for spi.
then what would happened if copy waza aplly correctly? the result will be about the same as this:
- 已加入
- 9/18/03
- 訊息
- 14,397
- 互動分數
- 2
- 點數
- 0
