The benchmarks I've (we've?) been waiting for...

[PetNorth]

Oh I see...You mean that hard to measure bullshite??? Dude, I have athlon xps in my house now and have had that 3200+ right next to my 2.4 at 3.2ghz and I don't know what the helll you are talking about....

Thats your problem, not mine.

I give give you hardcore numbers and real data not a sense...

LOL hehe hardcore numbers and real data? I posted hadcore numbers, real data and documented above of a hardware page. With them, your hardcore numbers and statements are nothing. I repeat them, perhaps you have not found out:

No lag!!! LMAO...you guys and your NAV in the background...Is it scanning the files or just active in the Systray!!!

Obviously it is scanning the files, all entire HD.
Obviously NAV active in the systray, no sense (at least in Athlon LOL ;-) ). 10 explorer open, no sense; 10 programs just active in systray, no sense... 7 programs open, no sense etc

Encoding and gaming just means I can game with about a 5-7% loss in max fps in gaming

Sure? about 5-7% loss fps only? sorry, but... LOL

Let's see: Flight Simulator main task, WME9 encoding in background (for the fps results, I supose with normal priority process):

http://common.ziffdavisinternet.com/util_get_image/5/0,3363,sz=1&i=57173,00.jpg

P4EE 3.4 running only Flight Simulator: around 55-60 fps avarage.
P4EE 3.4 running FS with WME9 encoding in background (I think without set less priority for encoding): 17fps avarage (with habitual drops to 7-8fps).

Is this result a 5-7% fps loss?
Is this avarage 17 fps with FS and WME9, playable?

Of course, in that situation, 4fps avarage for A64 in FS, encoding with WME9 in background without set less priority for encoding task, is absolutely unplayable. But, is playable a 17fps avarage with habitual drops to 7fps in P4EE 3.4? Obviously, either (at least if you want to play decently).

This is what I said above, with multi tasking tests in Xbitlabs: a real life multi tasking situation (tests 1 and 2) A64 is better. An irreal life multi tasking situation (test 3) P4C is better (as I said before, is less bad). In that test 3 situation *any* single CPU works decently. So, what's the win? where's the advantage?


Now, another example: Dungeon Siege main task, Sound Forge encoding in background (for the fps results, I supose with normal priority process for Sound Forge set to less than normal):

http://common.ziffdavisinternet.com/util_get_image/2/0,3363,sz=1&i=29026,00.gif

Well in this situation, playing Dungeon Siege and meanwhile encoding with Sound Forge in background (I supose, with priority process for SF set to lower than normal -for the high fps avarage obtained-) even AXP 3200+ obtains more fps than a P4EE 3,2. And A64 3200+ and FX-51 obtain between a 25 to 40% more fps than P4C 3,2 and P4EE 3,2.

Yes, P4 finish Sound Forge encoding in background sooner than Athlon, but really is that the objetive in this multi tasking scenario???????? or the objetive is playing game -our main task- smoothly possible and with great quality?????? Who's matter SF finish sooner??? I think, the answer is obvious. And I think is obvious which one is the more recomendable situation. So, I think it's unnecessary I answer this questions here, isn't it?

Priorities have little effect with HT as it seems to override everything. I played with each and every one of them and they usually seem to have negative or opposite effect. It appears HT schedules the threads much more efficiently....

It will be in P4. In Athlon is absolutely great, and it lets you *really* to continue working (or playing) normally, even doing some hard work in background.

Regards.

 

[OddTSi]

This all comes down to plain and simple hypocrisy. When Intel execs as well as people on this board say that 64-bit processors are not NEEDED on the desktop, AMD-fanboys (most recently PetNorth in the thread in GH) are quick to respond by saying that if we just used what was needed we wouldn't be advancing technologically so it's good to have advanced technology even though most of us won't have a need for it.

Now that someone posts multitasking benchmarks (something that I've been wanting AnandTech to do; since they're posting 64-bit performance benchmarks for the AMD-camp, why not be equal and post benchmarks showing the benefit of HT) that show the performance improvement of HT enabled chips, AMD-fanboys are quick to say, who cares if it makes things faster, most people don't need it and won't use it often enough.

Bunch of hypocrites.

 

[PetNorth]

You don't read what others write. I've said P4 NEEDS HT abolutely. And HT IS GOOD for P4 architecture. Why don't your read, Intel-fanboy?

 

[Duvie]

Originally posted by: PetNorth
You don't read what others write. I've said P4 NEEDS HT abolutely. And HT IS GOOD for P4 architecture. Why don't your read, Intel-fanboy?

No!!! What you seem to keep reiterating is that a p4 without HT is somehow subpar to the AMD athlons...That is plain and simple not true....

P4 does not need HT!!! In many apps when I shut it off I was still faster at 3.2ghz then the barton at 3200+...With it on it flat out killed it and made it run in some test worse then my 2.4c at default....

It is clear your AMD fanboy glasses hampers your vision....

By the way I will be running the flight sim 2002 here in a bit when I load it up...I only have that game, CSI (which is not much of test) and the UT2003 which is more then enough....

 

[Duvie]

I read your pics and I hat to throw this word out but I think you are a moron!!!


Did you even look at them??? PLus go ahead and give us the rest of this link so we can see things such as set up and other specifics to the applications.

The P4 is obvious loses to a A64 in games as this is no surprise...Are you color blind??? If the multitasking is true the A64 barely got near the p4's lowest frame....Did you look at the colors right???

The the second one....Cmon, we all now the barton 3200+ was still competitive in gaming versus the p4 3.2c it was just at the end of the platform for speed. Therefore had they shown us the times of each app singular you would see what I have said all along. The game in the foreground on the axp will command all the power and the encoding will get much less cycles when the game frees it up. NOw we can figure that the p4 will still drop its encoding time by 25-45% depending on the app. Yet it still beats the AMD by 68%...PLus I don't believe soundforge is not HT enabled and therefore probably only had a 10-15% max lead on the AMD in single app speed....

I think you don't even how to enduce any conclusions from this....


I will rerun and post my UT2003/encoding + FS2002/encoding and leave it at that...I don't find you worthy of discussing this with anymore. I see that in other threads this is just how you act and are...


UPdate: Testing...


OK first thing with system at 3.5ghz as below and vid card set to optimal quailty through my cat drivers and set to 1280x1024/32bit with HIGH quality settings in scenery, aircraft, antialiasing on, extended texturing, bilinear filtering....also set refresh rate to 85

This game even by itself according to my cpu usage chart does not take much cpu usage...It showed an average of 20-30% with maybe a max and think that was stopping and minimizing the app of 35%.....So I am not seeing how FS2004 is that much more taxing. In this day and age the game cards are very important in handling this...


Fraps tells me

FS2002 (single) = minimum of 5 minutes = max 63 average 58 minimum 1 (startup)

TMPGenc (multithreading on in app) = 2:08

TMPGenc (multithreading off in app) = 2:41 (diff vs on = +26% in time)

________________
FS2002 (multi) = I ran it until I heard the sound of TMPGenc ring = max 61 average 54 minimum 1 (startup) (-7.4% in avg fps)

TMPGenc (multithreading on in app) = 3:43 (+74.2%)

__________________
FS2002 (multi) = I ran it until I heard the sound of TMPGenc ring = max 63 average 56 minimum 1 (startup) (-3.5% in avg fps)

TMPGenc (multithreading off in app) = 4:19 (+60.8%) (vs on 16.1%)


for food of thought...

Barton 3200+ TMPGenc (single)= 3:16 (54.2% more then 3.5 w/ multithreading on)

p4 2.4c TMPGenc (single) = 3:05 (44.5% more then 3.5 w/ multithreading on)



The conclusion here is exactly what I sadi with gaming app obviously in the foreground since it takes the whole screen....

Other noticeable facts is that HT on in the app alone of TMPGenc has tremendous gains and I noted this quite awhile back in single mode vs off....When multitasking is going on this gap shrinks as you notice only a 16% gain....


I will do UT2003 next......



The more I think about it that test is obviousluy flawed....No way the gaming app in the foreground in either settings unless in the amd systems he made the game low priority or the encoding higher then normal would you get those results...I have conversed with other p4c users and I have yet to hear anyone have those type of degradation in gaming fps while doing cpu intensive apps like encoding....Just don't buy it....Find me another link...HT is still wise enough to know anything in the foreground will get priority since display is happening...



UT2003....

ut2003 (bombing run 1280x1024 with high settings throughout)

UT2003(single) = fraps avg 60 max 132 (30 minutes of play)

TMPGenc (single w/ multithreading on) = 18:08 (longer test clip)

UT2003(multi) with TMPGenc doing a much longer encoding so I can play a bit to work in the averages and big battles to stress the fps....................= fraps avg 56 max 128 (-7.1%)

TMPGenc (multi) = 30:50 (70% longer) I had to run my 2.4 at 12x166 to get it to take longer in single app mode with multithreading on...So therefore it ran as fast as a 2.0ghz p4c or since the HT helps quite a bit more like a 2.26-2.33ghz p4 w/o HT....


The 7% just like I said...I bet if I run w/ multithreading off in TMPGenc I would get the 5%..

Edit: The cpu usage for me with UT2003 is 50% on an HT system which means about 100% on a non HT system....

 

[PetNorth]

Prove with Windows Media Encoder 9, not with TMPGEnc, and tell me what happens with the game frames meanwhile WME is encoding in background with normal priority...

PS. I've never used TMPGEnc, but I've just prove it and I've noticed a curious thing that I don't understand why it happens: when I start encode, this program automatically drops priority process to "Low" ¿?¿? so I can continue with other tasks normally without any noticiable perfomance low, game, premiere, any program runs perfectily without noticiable loss. Of course encoding takes more time. I don't understand why this silly program drop priority to low automatically when encoding starts, without I manually do it. ????

PS2. About if P4 needs HT or not (needs it like we need the water), hehe well, obviously we will not agree at anytime.

PS3. I would be thankful to you that you were not a bad-mannered one and you retire your insults.

 

[chsh1ca]

Originally posted by: Duvie
This game even by itself according to my cpu usage chart does not take much cpu usage...It showed an average of 20-30% with maybe a max and think that was stopping and minimizing the app of 35%.....So I am not seeing how FS2004 is that much more taxing. In this day and age the game cards are very important in handling this...

Waste of time to even post your silly FS2002 benchmarks. You need to compare apples to apples. You don't see anyone going and doing a comparison of UT2K4s performance using UT2K3 now do you?

I will do UT2003 next......

The more I think about it that test is obviousluy flawed....No way the gaming app in the foreground in either settings unless in the amd systems he made the game low priority or the encoding higher then normal would you get those results...I have conversed with other p4c users and I have yet to hear anyone have those type of degradation in gaming fps while doing cpu intensive apps like encoding....Just don't buy it....Find me another link...HT is still wise enough to know anything in the foreground will get priority since display is happening...

Actually, I'd be VERY amused and surprised if Intel could make processors that did the operating system's job.

Ever since you did these benchmarks you've abandoned reason. WTF is comparing FS2002 to FS2004 and saying they would have equal performance about? You don't even own FS2004 to test with. You also ignored the other key part of the test which was WME9 encoding in the background. You say you will prove the numbers wrong, then go on to prove an entirely different set of numbers wrong. I'm beginning to think you are becoming one of the zealots you love to bash.

Edit: The cpu usage for me with UT2003 is 50% on an HT system which means about 100% on a non HT system....

Oh man, now this, this seals the deal for me. Newbies get flamed for this kind of idiocy, just keep that in mind when someone busts out a flamethrower and calls you a troll.

 

[Keysplayr]

Synonym for AnandTech: Testosterone -R- Us.

 

[mechBgon]

<--- likes chocolate pudding.

 

[DAPUNISHER]

Originally posted by: mechBgon
<--- likes chocolate pudding.

The Jello brand or the kind you eat on your MTB after the snow thaws?

 

[Duvie]

chsh1ca (shite talker)

With HT enabled it does in fact take priority settings out of the mix...I tested this theory a few weeks back. Changing them often had little to no effect but often did the opposite of what you would think the channing of the priority would do. With HT off it worked fine...

SO I SAY!!!!

Get yourself a P4c you shite talker and try this stuff out for yourself. NO matter what I test you argue everything yet I fail to ever see you test anything to disprove me. SHITE OR GET OFF THE POT....



I have had FS programs since 3.1 days and unless somehow windows went night and day from one version upgrade then the usual update the city textures and add a few more planes to fly their should have been no revolutionary advances since the minmum system required is not all that different from the FS2002...There is as far as I know no 2003 version so this is 1 step up...

I will run windows media encoder since I believe I downloaded it to run pcmark2004. I will see but since I don't have FS2004 or soundforge I guess I shouldn't bother, huh???

The point is the theory of it. Why would I run 2 different test of games and have similar results??? Why when in the past the both the p4 w/ HT off and the Barton 3200+ act quite similar and the P4 with HT did the game with less fps (albeit the numbers say not that bad) but completed the second app much faster. I brought this up awhile back and said if the game is still near same levels but did the 2nd app considerably faster (faster then the fps decreased in terms of percentage) why wouldn't it be considered a success....


Last but not least!!!! Cause I am done with talking to the likes of you who question all that is before them with no experience with this system....Dont by any means think I haven't confirmed these numers with other p4c users...Maybe they are smarter and choose not to mention them in threads so they don't get attacked in the threads and basically called a liar....

Take programs like

SETI
F&H
TMPGENC (with multithreading off)
BESWEET
POV-RAY
Prime95

All these apps on an HT system will max out in single mode as 50% in the task manager with 2 cpu charts open...This is fact shite talker....

Now take these apps and disable HT in bios and cpu chart will show only one chart and these same apps will be maxed at 100% in single mode....


I know cause I own one and you obvious with your lack of correct knowledge do not....

 

[chsh1ca]

Originally posted by: Duvie
chsh1ca (sh!te talker)

With HT enabled it does in fact take priority settings out of the mix...I tested this theory a few weeks back. Changing them often had little to no effect but often did the opposite of what you would think the channing of the priority would do. With HT off it worked fine...

HT is still wise enough to know anything in the foreground will get priority since display is happening...

Amusing.

SO I SAY!!!!
Get yourself a P4c you sh!te talker and try this stuff out for yourself. NO matter what I test you argue everything yet I fail to ever see you test anything to disprove me. SH!TE OR GET OFF THE POT....

So you are ultimately right about the processor knowing WTF is going on with other parts of the PC, or which threads should be prioritized how (which is actually done in the operating system) because you own one? I was simply saying compare apples to apples. You claim the FS2004 benches are bullshit by benching FS2002. LMFAO.

I have had FS programs since 3.1 days and unless somehow windows went night and day from one version upgrade then the usual update the city textures and add a few more planes to fly their should have been no revolutionary advances since the minmum system required is not all that different from the FS2002...There is as far as I know no 2003 version so this is 1 step up...

Yet you don't know for sure. In your own words:

Get yourself a [copy of FS2004] you sh!te talker and try this stuff out for yourself.

I will run windows media encoder since I believe I downloaded it to run pcmark2004. I will see but since I don't have FS2004 or soundforge I guess I shouldn't bother, huh???

Since you claimed the benchmarks were complete BS, yes, back up what you said. I didn't talk sh!te about anything, I merely pointed out the rather obvious flaws in your test as a comparison to another test USING COMPLETELY DIFFERENT APPS.

The point is the theory of it.

Bull. If it was just the 'theory of it' then don't come out and say:

This game even by itself according to my cpu usage chart does not take much cpu usage...It showed an average of 20-30% with maybe a max and think that was stopping and minimizing the app of 35%.....So I am not seeing how FS2004 is that much more taxing. In this day and age the game cards are very important in handling this...

Back your wild accusations up is all I'm saying. If anyone tried to release a legitimate estimation of UT2K4 performance by using UT2K3 they'd be laughed out of benchmarking by the people who have half a clue. Instead of making the gross assumption it is identical, why don't you find out? That you didn't bother and seem to be under some delusion regarding how things work (the foreground process is not necessarily the highest priority process, and YES, that can be configured) lead me to believe you don't know as much as it would seem. The P4C is a great processor, sure, but it is not the holy grail of multitasking you imply with your 'benchmarks'.

Why would I run 2 different test of games and have similar results???

2 whole tests? That it? Come on man, if you are gonna come out and call another guy's benchmarks bs you should back it the hell up with benchmarks of your own.

Why when in the past the both the p4 w/ HT off and the Barton 3200+ act quite similar and the P4 with HT did the game with less fps (albeit the numbers say not that bad) but completed the second app much faster. I brought this up awhile back and said if the game is still near same levels but did the 2nd app considerably faster (faster then the fps decreased in terms of percentage) why wouldn't it be considered a success....

Last but not least!!!! Cause I am done with talking to the likes of you who question all that is before them with no experience with this system....Dont by any means think I haven't confirmed these numers with other p4c users...Maybe they are smarter and choose not to mention them in threads so they don't get attacked in the threads and basically called a liar....

Who called who a liar? I suggest you brush up on your english bud. I question the validity of statements made by people who are knowledgeable enough to be dangerous and make incorrect statements like:
The cpu usage for me with UT2003 is 50% on an HT system which means about 100% on a non HT system.
HT is still wise enough to know anything in the foreground will get priority since display is happening.
Ask anyone who has done ANY kind of threaded software development, they can clue you into why these statements are wrong since you aren't "talking to me".

Take programs like
SETI
F&H
TMPGENC (with multithreading off)
BESWEET
POV-RAY
Prime95

All these apps on an HT system will max out in single mode as 50% in the task manager with 2 cpu charts open...This is fact shite talker.

Lol, now your unequivocal explanation is that task manager says it's 50%? LOL, that's too funny. The reason task manager shows two is because it believes it is a DP system, it doesn't have a way of accurately measuring processor usage since it is not aware (and indeed, it would be INCREDIBLY difficult to make it aware) of the fact that the "two" processors it sees are sharing execution units.

Now take these apps and disable HT in bios and cpu chart will show only one chart and these same apps will be maxed at 100% in single mode.

Your flawed understanding of the system is indeed not my fault. Instead of simply "using" the system, you should brush up on the actual design and architecting of the processor, with a touch of understanding of threaded application development and how windows sees processors. Just a suggestion since you are way off on some of what you are saying here.

I know cause I own one and you obvious with your lack of correct knowledge do not.

LMFAO. Wait, that means I know EVERYTHING about my Honda Accord, I do own one after all. Oh, and I am the new local graphics card architecture and IQ expert for the GF3 series of cards, since I own one.

I just realised I'm breaking the rule in my sig.