Stress Test MK II

Page 24 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Yup I think they said HT was on again.

And was that begrudging acknowlegement I seen, that without HT the P840 got it's ass spanked on ALL 4 apps?
Anxiously now awaiting the english version of update 9.

<P ID="edit"><FONT SIZE=-1><EM>Edited by Weescotty on 06/17/05 07:43 PM.</EM></FONT></P>
 
The only thing I like to see with the single app test is the divx. But I dought that going to be setup right. IE seting up the program if possable to use 2 or 4 treads. but I gess we will soon see.
 
All man now there doing Lame encodeing. That is why more IO then cpu. They should have the same HD and CD/DVD drives. So what dose this test how much cpu the SATA and the IDE controller use. Hopefully they will get around to translating the 9th update so we all can laught.
 
Well if I may say that was one hell of a read or long thread. Nice too see these forums still discuss the facts for the most part.

It's somewhat uncomfortable or embarrassing watching thg try to get the outcome they desire.

Hopefully Tom and family are having fun sipping pinacolodas or something in the Caribbean, Mexico or Thailand whatever. All the best to you Tom you earned it, ya spoke your mind and put all the cards down on the table, was truly a great fun website to read. However, I’ll bet it makes you cringe a little reading the new and improved thg. Ahh have another margarita it’s not your problem anymore.

I am reasonably confident Tom was handsomely compensated to step out of the picture. Would anyone here do less or more. Not likely the almighty buck dictates what we do for the most part and I would be no different. BTW Tom taught me how to overclock my first cpu was very exciting at the time I felt like I was some kind of rocket scientist and when it really worked it was total wow. Celeron 300mhz to 450mhz up 50% boost for nothing at all price wise.

Remember when Tom left and Omid entered the picture and accused the world or the idiot world of being amd fanboi's. Does anyone here remember that column? For all I can remember Omid coined the phrase fanboi based on that column, OH the irony in it.

Anyone know if Omid still writes for thg editor in chief or did the get rid of him?

I don’t know who runs thg but it’s defiantly time for a change.

IMHO, Anyway.



If I glanced at a spilt box of tooth picks on the floor, could I tell you how many are in the pile. Not a chance, But then again I don't have to buy my underware at Kmart.
 
What there an error mesage on the amd machine. Somebody do something intel for sure to win lame encodeing now. See no machine is 100% perfect. Any way there stil more to come and update 9 well it was ok stil did not make sense in some area, and stil showing THG German most not have that full understanding of HT and windows piorty. oh well again
 
yes something does seem weird, I mean for the AMD system to beat the Intel system in both tests with and without hyperthreading on and then to lose when that is the only application running. seems fishy to me.
 
I bet your proud of yourself :lol: Has this thread broken any THGC records yet?

<font color=green>NED FLANDERS FOR PRESIDENT</font color=green> Its justa nother gansta PARTY!
Intel P4 Extreme(3.73)@<font color=green>5.6Ghz</font color=green>
Asus P5AD2-E-Premium
Crucial Ballistix DDR2 667@DDR2 855
 
Wow they fix the AMD and now it ahead again. And look divx looks like it is useing all four of the Intel logicail units. To bad AMD is in the lead again. Hopefully they will tell us what the error was on the AMD Machine. I think it just was the program to send the stats to the server, because network traffic stoped for that time. It allmost over.
 
Intel seems to have problems with hyperthreading, it can't utilize both CPU's well enough.
<A HREF="http://www.tomshardware.com/stresstest/load.html" target="_new">http://www.tomshardware.com/stresstest/load.html</A>

But if with HT on those average loads should be 50/50, it still doesn't explain current balance.

Perhaps Intel is suffering high latencys and can't utilise it's bandwith.

<font color=red>"We can't solve problems by using the same kind of thinking we used when we created them."
- Albert Einstein</font color=red>
 
I to was supprised to see the X2 smoking the P-EE 840 by that much in DivX, but I noticed the loads on the AMD system (from what I have seen at least... they really need a runninng graph) average in the 90+% for both CPUs, but the Intel system does not load that well. Most of the time I see 60-80% loads and usually one CPU is about 10-20% lower than the other.

So far AMD has beat the P-EEin every teswt and I think FarCry will be no differant. So the X2 wins hands down in ease of building a stable system, power draw, and single and dual application loads and the P-EE does really good in DivX as long as there are 3 other applications loading the CPU.... Humm. I wonder.

One telling thing though is that AMD's X2 are alvaluable in pretty good quanities, but they are selling like hotcakes. Places that said they recieved large quantities are already out. AMD has a hit and these things are flying off the shelves. Considering they charge such a primium over what Intel is going to see their P-Ds for it really shows what people think of them. This should help AMD gain market share and increase proffits.
 
AMD's X2 are alvaluable in pretty good quanities, but they are selling like hotcakes. Places that said they recieved large quantities are already out.
Good move from AMD to keep those prices above 500, they make some much needed cash.
Also im not surprised that X2 is selling well, it is after all a CPU that is for enthusiast and for people that know this sh**.
It does give good value for the money.
Just compare it to +3800-4000 models, price is same or 20% higher for 4400 X2, so no point in buying single core.

<font color=red>"We can't solve problems by using the same kind of thinking we used when we created them."
- Albert Einstein</font color=red>
 
I'll fix another wrecked link of you ...

<A HREF="http://www.tomshardware.com/cpu/20050603/index.html" target="_new">UPDATE 9</A>

<font color=red>Sig space for rent. make your offer.</font color=red>
 
I must make this post or both computers crash..

<A HREF="http://koti.welho.com/pnystro2/somepics/thgc.postcount.PNG" target="_new">I saved us..</A> :lol:

<font color=red>"We can't solve problems by using the same kind of thinking we used when we created them."
- Albert Einstein</font color=red>
 
Did you notice the point where that pic ends ?


Maybe im too mean, im starting to feel bad.. :frown:


yeah right.. :evil:

<font color=red>"We can't solve problems by using the same kind of thinking we used when we created them."
- Albert Einstein</font color=red>
 
The German site is sometimes days ahead. They had the #9 update several days before. The AMD system is just pulling farther ahead and the X2 is going to be a great seller. Now that AMD has split off the memory division, they should show a good proffit for ther CPU division. I think they should split off the Geode too as I do not think it will ever make them money. Heck they could take a Semperon and run it at say 1Ghz full load and 200Mhz idle. It could do .8 and .5v respectively. I know my Athlon64 3500+ Winchester core runs at like 10w idle at 1.1v and 1Ghz with Cool and Quiet and tops out around 35w or so at the full 1.4v and 2.2Ghz. I also know this thing at 1Ghz can handle playing a DVD and still run at about 5% load. It could likely handle almost any games too at 1Ghz.

There is also a nice rumor (that Anand is even looking at) that puts a DX9 LongHorn capable integrated graphics and I/O controller on chip sometime next year. Think of a 65nm dual cored Athlon64 X2 with a cut down x300 core with one pipeline. You'd think it would be slow, but think of it linked to a dual channel DDR2 667-800Mhz and a core running at 2Ghz or faster. Now think that it would only add about 20% or so to the die surface making it a cheap add-on. Now think of how little it would increase the power on this chip and how much power it would save by removing the north and south bridge. Now also thing of the reduction in size on boards for SFF computerss and laptops. Also think of the reduction in costs to make the boards. Even 1 pipeline at 2Ghz would be fast....

Now I know it would likely not be an ATI or nVidia graphics core, but I know VIA is really down on its luch and SiS and ULi are not doing so well either. They all have some sort of DX9 graphics either out or coming out and they do not perform too well. Even discrete cards cant keep up with an x300 SE or 620 TC card, but think now of taking that core and running it at 4-5 times the speed. Think also of increasing the memory bandwith the same or more. No it would not beat an x700Pro or 6600GT, but it might come close to say a x600Pro and 5700U. Now think that this would likely save $20-50 of the chipset and motherboard costs and would liekly only increase the cost of the CPU by $10. Also figure this would be a reduction of power for the chipset of likely 20-30w. In a Thin and Light market it might actually draw less power than a Yonah as the chiposets right now kill the Turion's battery performance. This would also be a great server chip and perfect for blades. For the full sized desktop it might not help much, but if you can use a PCI-e x16 card and this then you still might be able to make good use of it to drive a second or third monitor.

The other rumor that I am kind of interested in is AMD might add co-proccessors again to some of their chips in 2007. The cool thing is they are already looking at PCI-e 2, DDR3, HT-2 bus and quad cores. Now think of adding a dedicated tiny special purpose logic unit. It could either been an SSE4 unit to help boost HD-DVD decoding speeds or there is also talk that it could a a logic unit to help split tasks between the cores including non-threaded applications. That would be VERY mighty. Instead of being like a Cell chip with one big core splitting the load to 8 smaller ones, it would be one small one splitting the load between 4 larger ones... 4 2+Ghz cores with 1MB L1 cache with what nine exicution units. The Cell cores are likely going to be clocked faster at 3.2Ghz with only have 256k of cache and only 2 exicution units.

Intel is not sitting still either, but their only cool core looks to be a quad cored P-M with an integrated memory controller and a 4MB shared L2 cache. This is going to be a XEON chip actually and might use a but based on an advanced Itanium2 bus (and possibly use the same actual boards). Nothing else Intel has coming up looks even close.
 
Great post.

AMD just needs their own Dell.

Maybe i'll start to make computers. :smile:

<font color=red>"We can't solve problems by using the same kind of thinking we used when we created them."
- Albert Einstein</font color=red>
 
lol, I don't think it has - in the polls section there is a thread with over 70 pages 😱

I <b>dare</b> you to read it from start to finish. :lol:

______________
<font color=green>NED AND MOZZARTUSM - REAL (P)RESIDENTIAL CANDIDATES</font color=green>
 
I know my Athlon64 3500+ Winchester core runs at like 10w idle at 1.1v and 1Ghz with Cool and Quiet and tops out around 35w or so at the full 1.4v and 2.2Ghz.

...you're chip runs at 67W +/- at full load I belive. Where are you getting your values from?

______________
<font color=green>NED AND MOZZARTUSM - REAL (P)RESIDENTIAL CANDIDATES</font color=green>
 
Well IMO the 9 th thg update was the worst one yet. Regarding intel biased wording :S.

Anyways nothing has changed.

If youre spending $$$ on hardware you should allways research and buy according to your own specific needs.

PS: and dont worry when it's all outdated after a year 😀 Happens to all of us 😀
 
Have they disabled Hyperthreading again??
Look at the CPU load graphs and it's saying hyperthreading disabled with 50% CPU utilization on each CPU?

The only thing each of these tests has proven (other than the original stability issues with intel), is how the schedualer works on XP and Hyperthreading, i.e. a low priority thread can negativly effect the performance of higer priority threads.
Please can they run the Test with HT switched on and 4 threads of the same priority. Only then can we see the true performance benefits of Hyperthreading in a dual core environment, and only then can we see the performance differences between AMD and Intel.
 
It seems that Intel can't utilise both CPU's 100%, propably because it cant use all the available bandwith (high latency).
Or it also can be a HT issue, but i doubt that.

But im more worried about <A HREF="http://koti.welho.com/pnystro2/somepics/balance.PNG" target="_new">this, which i see quite often.</A>
AMD has no problems filling it's pipelines, so how divx can only take 50% ?
It doesn't happen all the time, but it does happen quite often.

Maybe it finishes a bunch of frames, then when waiting new ones to be loaded it's CPU usage is lower.. ?

Or are they playing games with their priority controlling app ?

<font color=red>"We can't solve problems by using the same kind of thinking we used when we created them."
- Albert Einstein</font color=red>
 
Maybe it's doing something similar to DVD shrink that I run regularly at home, it spends alot of time doing some work, then it fills it's buffer area's that don't take too much CPU utilization (i.e. it reads the disk, etc,etc).
With modern drives the Southbridge can generally read the disks with little CPU involvment