+i/ea2GJLXi5f2Q-PC HELP from China
qiang gao
Jian Yu wrote:
>
> Something to consult to you: my computer is a IBM 兼容机,
> cyrix 486 133mhz, 内置 BIOS (Award), ISA Bus (PC/AT),
> VGA (显示卡 S3 Trio 868). Today I wanted to
> upgrade its speed by means of re-arrange the jumpers
> on the main board. According to the main board manual,
> I set the jp28 as 1-2 off, 3-4 and 5-6 on.
> By this the speed of the CPU was supposed to be raised to 160 mhz.
> I did succeed for one time. When the re-setting of
> the jp28 was done and I turned on my computer again,
> self-test printout showed that the CPU speed jumped up to 160.
> And Norton Utilities test reported that the CPU performance
> index was raised from 297 to 350. That was exciting--but
> when I turned off my computer (to put back the cover and
> other stuffs), then turned it on the next time, it failed to
> boot in. The screen just didn't show anything.
> Blank at all. And the keyboard was dead, too. Hard disk drive
> seemed to got stuck there because its light kept
> lighted. I tried for several times, for one time the screen
> gave the picture of the RAM test, then freezed right there,
> before MS-DOS started.
>
> Based on this information would you give me some hints
> what the problem could be? (When I switch the jp28 back
> to its original status, the computer worked O.K. again. So
> nothing serious happened to the hardware.)
>
> /Shisheng
If that computer is supposed to work at 160Mhz correctly, why would
its maker selll it as an 133mhz machine and throw some bucks into
water? Call them to find it out why?
Many things can be wrong(based on the description). Since you have an
ISA machine, it could be your mother borad cannot handle frequency
multiplied 160Mhz CPU, or it could be a secondary cache problem. Or it
could be you need a upgraded system BIOS. Or it could be your CPU
fun/heat sink is not sufficient at 160Mhz. You are asking for troubles
when you are beating your computer too hard :-)
Quanyi Sun
qiang gao (g...@ece.arizona.edu) wrote:
: Jian Yu wrote:
: >
The main thing is that you MUST be sure your bother board supports
160 Mhz. If so, 133 MHz -> 160 MHz is still too hard. Successful stories
I have heard is: 486INTEL 33MHz -> 50MHz(not 66MHz), P100MHz -> 120MHz.
It seems a 20MHz would be a max jump. After that, even your computer
boots correctly, your MS-Windows may crash.
Kevin Wang
Basically it is recommended not to haul CPU speed in this way,
but sometime people do things like that. I am not sure about
cyrix's one, lots of people in AMD do boost the 5x86
(486 133Mhz version) to 166MHz, and systems run fine. Maybe
cyrix's one doesn't really provide such performance margin.
One other thing is, like you said, the system ran fine the first
time. This may be caused by the CPU temperature. When temperature
goes higher, CMOS circuits usually get slower, meaning
it cannot keep up with the clock ratio you set up. And I also
know cyrix's very hot. I think your setup is just at their
margin edge, try to lower the clock ratio to 150Mhz, leave
the cover open, and change a better heat sink and cooling fan
-- if you really want to upgrade sys this way...
Wish it may help.
qiang gao
Quanyi Sun wrote:
>
>
> The main thing is that you MUST be sure your bother board supports
> 160 Mhz. If so, 133 MHz -> 160 MHz is still too hard. Successful stories
> I have heard is: 486INTEL 33MHz -> 50MHz(not 66MHz), P100MHz -> 120MHz.
> It seems a 20MHz would be a max jump. After that, even your computer
> boots correctly, your MS-Windows may crash.
Not really, the motherboard does not have to support 160Mhz(very few
backplanes can support 160Mhz directly). Since
it is a clock multiplied CPU, it is 160Mhz internally. If the cache
speed and hit rate is good enough, and loading is not a problem,
it would still work. For an ISA bus, line loading is difinitely a
problem, though.
The examples you give are not for clock-multiplied CPUs, so they are
different.
qiang gao
Kevin Wang wrote:
>
> Basically it is recommended not to haul CPU speed in this way,
> but sometime people do things like that. I am not sure about
> cyrix's one, lots of people in AMD do boost the 5x86
> (486 133Mhz version) to 166MHz, and systems run fine. Maybe
> cyrix's one doesn't really provide such performance margin.
>
> One other thing is, like you said, the system ran fine the first
> time. This may be caused by the CPU temperature. When temperature
> goes higher, CMOS circuits usually get slower, meaning
> it cannot keep up with the clock ratio you set up. And I also
slower? Complementary MOSFET does not do that before your chip is melt.
It becomes noisier and turns zeros into ones and crashes your computer.
Benny
Sorry about the mess. I saw this guy's post b4, but forgot
to answer it right away. Anyway, here's my point.
What you did is called overclocking, which is a dangerous
way to speed up your machine. For a Intel CPU, you should
be able to go a little bit high,( Is it worth?) but you leave
your CPU in a very risky place. This kind of speed-up causes
over heating which is a cause of the major faliue of the CPU.
Try to change your setting back, if it is still working, then
u r lucky. If not, just pray bah.
Kevin Wang <kevin...@amd.com> wrote:
>Basically it is recommended not to haul CPU speed in this way,
>but sometime people do things like that. I am not sure about
>cyrix's one, lots of people in AMD do boost the 5x86
>(486 133Mhz version) to 166MHz, and systems run fine. Maybe
>cyrix's one doesn't really provide such performance margin.
Benny
风吹旗动。
“是风在动。”
“不，是旗在动。”
“是你们的心在动。”
滚滚长江东逝水，浪花淘尽英雄。是非成败转头空：青山依旧在，几度夕阳红。
白发渔樵江渚上，惯看秋月春风。一壶浊酒喜相逢：古今多少事，都付笑谈中。
Weining Wang
********************************************************************.
For your info, go to
You should get a clear idea about CPU's overclocking.
WWN
--
*************************************************
* Weining Wang | E-mail: wwei...@wpi.WPI.EDU *
* | wwei...@ziplink.NET *
*************************************************
qiang gao
Weining Wang wrote:
>
> In article <3303A6...@ece.arizona.edu> qiang gao <g...@ece.arizona.edu> writes:
> >Quanyi Sun wrote:
> >>
> >
> >>
> >> The main thing is that you MUST be sure your bother board supports
> >> 160 Mhz. If so, 133 MHz -> 160 MHz is still too hard. Successful stories
> >> I have heard is: 486INTEL 33MHz -> 50MHz(not 66MHz), P100MHz -> 120MHz.
> >> It seems a 20MHz would be a max jump. After that, even your computer
> >> boots correctly, your MS-Windows may crash.
> >
> >Not really, the motherboard does not have to support 160Mhz(very few
> >backplanes can support 160Mhz directly). Since
> >it is a clock multiplied CPU, it is 160Mhz internally. If the cache
> >speed and hit rate is good enough, and loading is not a problem,
> >it would still work. For an ISA bus, line loading is difinitely a
> >problem, though.
> >
> >The examples you give are not for clock-multiplied CPUs, so they are
> >different.
>
> ********************************************************************.
> For your info, go to
>
> http://sysdoc.pair.com/
>
> You should get a clear idea about CPU's overclocking.
>
> WWN
Thank you very much. Why don't you teach me something about it?
Is http://sysdoc.pair.com better than http://www.intel.com ? It
never fails to amaze me what computer junkies can come up with.
BTW, since you have a good idea about this overclock stuff, maybe you
can help me. I have three volumes of P5 technical data books to get rid
off for $100 OBO(there were $350 when brought new), leave me a note if
you know someone who might be interested. It may help them out when they
actually play with a CPU or two, "over-clocking" or not?
BTW, it helps to know that there is no such thing in engineering terms
as "over-clocking", Computer junkies invented it.
Below is what Intel says about "over-clocking". They should probably add
a clause "when used with a PCI bus".
----------------------------------------------------
Q13:
Will a remarked, or "overclocked", processor operate at the higher
speed?
There is no way of telling, as such a processor is being operated
outside the specifications for which we test the part;
the processor is being stressed beyond its guaranteed operating
limits. The results of this are unpredictable; errors may
occur, some of which will be obvious and others may be more
subtle, such as wrong answers for some calculations.
As a result, the processor may be operating incorrectly even when
it is not apparent.
In addition you may shorten the life of the processor by
overclocking. An overclocked processor may appear to work
correctly for several months and then begin to exhibit signs of
failure.
Dagang Wang
qiang gao <g...@ece.arizona.edu> wrote:
>Jian Yu wrote:
>>
Most likely, the fan/heat sink is not powerful enough for the cooling
of the CPU. Get a new CPU/MB. It's very cheap now.
ghost
wwei...@femme.WPI.EDU (Weining Wang) worte:
>For your info, go to
>
>http://sysdoc.pair.com/
多谢网友介绍。
在下原对超频持不信任态度，对此网页
详加研〔yan2〕究后毅然决定向虎山行。先上街
买了个风扇给中央处理器装〔zhuang1〕上 (原先只
有一个大散热器 )，开机一试，效果比
原来强了不少。现已将奔腾一三三超频
到一六六使用，到现在两个多小〔xiao3〕时没有
任何问题，速度没有用什么BENCHMARK
之类〔lei4〕测，但明显感觉快了。用手摸散热
器，还没有超频前只用散热器时热。
决定先试用一个星期，如无问题就试一
下超频到二百。
Hmm,前两天还想买一块一六六把一三三
换了，这下，十五刀搞定！
一定要用风扇，我买的ball bearing
fan with heat sink是很平常的，效果
比大散热器强多了。
qiang gao
ghost wrote:
>
> wwei...@femme.WPI.EDU (Weining Wang) worte:
>
> >For your info, go to
> >
> >http://sysdoc.pair.com/
>
> 多谢网友介绍。
> 在下原对超频持不信任态度，对此网页
> 详加研〔yan2〕究后毅然决定向虎山行。先上街
> 买了个风扇给中央处理器装〔zhuang1〕上 (原先只
> 有一个大散热器 )，开机一试，效果比
> 原来强了不少。现已将奔腾一三三超频
> 到一六六使用，到现在两个多小〔xiao3〕时没有
> 任何问题，速度没有用什么BENCHMARK
> 之类〔lei4〕测，但明显感觉快了。用手摸散热
> 器，还没有超频前只用散热器时热。
>
> 决定先试用一个星期，如无问题就试一
> 下超频到二百。
妈呀！ 你要领导第二次计算机革命？
风扇+大散热器。
我还可以教你一招。 从你们家水龙头接一
小管到散热器上， 没确你就可以试一
下超频到五百！
Weining Wang
In article <33064929...@news.oit.unc.edu> sorry, please post to newsgroup writes:
>wwei...@femme.WPI.EDU (Weining Wang) worte:
>
>>For your info, go to
>>
>>http://sysdoc.pair.com/
>
>多谢网友介绍。
>在下原对超频持不信任态度，对此网页
>详加研〔yan2〕究后毅然决定向虎山行。先上街
>买了个风扇给中央处理器装〔zhuang1〕上 (原先只
>有一个大散热器 )，开机一试，效果比
>原来强了不少。现已将奔腾一三三超频
>到一六六使用，到现在两个多小〔xiao3〕时没有
>任何问题，速度没有用什么BENCHMARK
>之类〔lei4〕测，但明显感觉快了。用手摸散热
>器，还没有超频前只用散热器时热。
>
>决定先试用一个星期，如无问题就试一
>下超频到二百。
>
>换了，这下，十五刀搞定！
>
>一定要用风扇，我买的ball bearing
>fan with heat sink是很平常的，效果
>比大散热器强多了。
>
>
Ghost,
I think there is another thing you need to do.
Get some silicon compound from Radio Shack.
The official name is Heatsink Compound, Cat. No.
276-1372, only $1.99. Put a little bit(a thin layer)
between the top of CPU and bottom of heatsink.
It will improve the heat transfer from CPU to
heatsink.
If your motherboard can support 75mhz(ASUS p55t2p4,
or Mustang R534), you can overclock your p133 to
p150 by setting 75x2. Since CPU runs at 75Mhz, it
will blow a real p150 even p166 into water. Try to
download some benchmark program to do the test.
Good luck.
WWN
ghost
qiang gao <g...@ece.arizona.edu> worte:
>妈呀！ 你要领导第二次计算机革命？
二百的PENTIUM又不是没有。
你应该〔gai1〕去了解一下芯片的制作。
另外你似乎根本没去 sysdoc.pair.com
看过。
>风扇+大散〔san4〕热器。
>我还可以教你一招。 从你们家水龙头接一
>小管到散热器上， 没确你就可以试一
>下超频到五百！
还用你来教？早想〔xiang3〕过了，不过在小管的制作
和连接上还有〔you3〕些技术问题要解决 :-)
还有一看你就是个计算机盲，CPU 在什么频
率下跑完全取决于妈妈板给它什么样的频率，
要到五百你得先给我特制一个妈妈板能提供
五百兆的频率。我那最多到二百，没法试超
到五百什么滋味。HMM，一定酷死了。
你要是能把这小管和妈妈板都整出来，一定
能赚钱。只是别弄得向银河那么大 (那可真
是用管子冷却的)， 那就不是PC了。
qiang gao
ghost wrote:
>
> qiang gao <g...@ece.arizona.edu> worte:
>
> >妈呀！ 你要领导第二次计算机革命？
>
> 二百的PENTIUM又不是没有。
> 你应该〔gai1〕去了解一下芯片的制作。
哈哈． 我应该去了解一下 SiGe 芯片的制作〔zuo4〕。
Si(CMOS，BiCMOS), 和 GaAs 的芯片都作过。
用您的技术， 二百的PENTIUM应该跑三百兆赫吧？
> 另外你似乎根本没去 sysdoc.pair.com
> 看过。
不错。
>
> >风扇+大散热器。
> >我还可以教〔jiao1〕你一招。 从你们家水龙头接一
> >小管到散〔san4〕热器上， 没确你就可以试一
> >下超频到五百！
>
> 还用你来教？早想过了，不过在小管的制作
> 和连接上还有些技术问题要解决 :-)
没听说过 LCM ？
>
> 还有一看你就是个计算机盲，CPU 在什么频
> 率下跑完全取决于妈妈板给它什么样的频率，
你知道不少计算机的东东嘛！
> 要到五百你得先给我特制一个妈妈板能提供
> 五百兆的频率。我那最多到二百，没法试超
> 到五百什么滋味。HMM，一定酷死了。
你不会换一个晶振器嘛？
回到“overclocking", 给你一点线索， 查一下"Weibull Distribution"
是什么一回事？
>
> 你要是能把这小管和妈妈板都整出来，一定
> 能赚钱。只是别弄得向银河那么大 (那可真
> 是用管子冷却的)， 那就不是PC了。
你出多少吧？ 比Pentium小一点点的ＬＣＭ．
ghost
qiang gao wrote:
> 哈哈． 我应该去了解一下 SiGe 芯片的制作。
> Si(CMOS，BiCMOS), 和 GaAs 的芯片都作过。
> 没听说〔shuo1〕过 LCM ？
> 你不会换一个晶振器嘛？
> 回到“overclocking", 给你一点线索， 查一下"Weibull Distribution"
得，还是〔shi4〕您牛，咱外行，看不懂您这些高级词，咱也
没做过 SiGe，GaAs，扯乎。
对了，overclocking也不止我一个试过，恐怕几十万
有了，您去给他们上几个关〔guan1〕于 SiGe, GaAs, LCM, 晶
振器、Weibull distribution 的 simenar 也许比造
LCM赚得快。
qiang gao
ghost wrote:
>
> qiang gao wrote:
>
> > 哈哈． 我应该去了解一下 SiGe 芯片的制作。
> > Si(CMOS，BiCMOS), 和 GaAs 的芯片都作过。
>
> > 没听说〔shuo1〕过 LCM ？
>
> > 你不会换一个晶振器嘛？
> > 回到“overclocking", 给你一点线索， 查一下"Weibull Distribution"
>
> 得，还是〔shi4〕您牛，咱外行，看不懂您这些高级词，咱也
> 没做过 SiGe，GaAs，扯乎。
我还以为碰到一大鬼呢 ． 咋的？
>
> 对了，overclocking也不止我一个试过，恐怕几十〔shi2〕万
> 有了，您去给他们上几个关于 SiGe, GaAs, LCM, 晶
这小〔xiao3〕鬼子挺喜欢搞大跃进。
Pao-Hung LIN
ghost wrote:
>
> 决定先试用一个星期，如无问题就试一
> 下超频到二百。
>
台湾人的坏毛病怎麽鬼兄也给染上了？
超频不论如何都会导至过热，用久了ＣＰＵ的那些众多接脚就容易断
只要断上一根，那就得重买ＣＰＵ了．啥时会断？你把ＣＰＵ拔出插入时
就是会断之时
不过人要有实验的精神［鬼我就不清楚了］
大不了明年可以买到便宜的ＭＭＸ ＣＰＵ
四匹