• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

C2D - Heavy multitasking performance.

E

enumae

Weaksauce
Joined
Jun 13, 2006
Messages
98
Hi guys, first thread, please be gentle...

I am looking to build a new system and I am a little worried about bandwidth.

Current rig for work is a Pentium D 820 over clocked to 3.4GHz, 2GB ram, and Abit AW8 motherboard, my system is pretty responsive, unless I am exporting a 36"x48" PDF in color then the system is a little slow in responding, it doesn't kill it, just a little lag.

I understand that K8 has more bandwidth than Conroe, in a heavily multitasking situation example...

Autocad 2007, Firefox, Thunderbird, Winamp, Frigate 3, and Adobe Photoshop, Acrobat, in case your wondering I have three monitors and there all doing something different, this is a daily norm.

My question is this, in a very heavily multitasking system does or would the bandwidth of K8 be better to support my needs than the Conroe?

Thanks in advance to everyone.
 
My question is this, in a very heavily multitasking system does or would the bandwidth of K8 be better to support my needs than the Conroe?
Click to expand...

It is clear enough that despite the poorer memory bandwidth, Core 2 Duo is the performance leader by a substantial margin at stock speeds.
Click to expand...

http://www.anandtech.com/memory/showdoc.aspx?i=2800&p=1

The article above basically shows that even though Conroe has (generally speaking) crappy bandwith, it still outpreforms the NetBurst and K8 in just about every application clock for clock.
 
You will never ever ever need more bandwidth than what the C2D can supply!

I'm sure that if you looked at your PDF program stats, while exporting one of those massive documents, you'd see a trend: High CPU utilization, High RAM utilization and lots of I/O writes.

The increased speed of the C2D processor will certainly help the "crunch" of the conversion. What's important with the RAM, is not it's bandwidth but the amount you have. There is enough bandwidth to do a task like the one you're describing and then some, as long as you have enough RAM in general.

Your and My computational needs are very similar: I'm always listening to iTunes while rendering 3D models, listening to lots of real-time effects in Cubase, while having IE on various pages. I have 2 monitors.

My X2 3800 (at 2.6 ghz) started to feel... a tiny bit slow during the highest overhead periods. My E6300 at 3.5ghz really doesn't show ANY lag at ANYTHING :p
 
EQTakeOffense said:
http://www.anandtech.com/memory/showdoc.aspx?i=2800&p=1

The article above basically shows that even though Conroe has (generally speaking) crappy bandwith, it still outpreforms the NetBurst and K8 in just about every application clock for clock.
Click to expand...

Thank you for the link.

I think I should have asked more about the FSB being a bottleneck in heavily multitasking scenarios vs Direct Connect, and HTT.

Would you still recommend Conroe?
 
Arcygenical said:
... There is enough bandwidth to do a task like the one you're describing and then some, as long as you have enough RAM in general...

My X2 3800 (at 2.6 ghz) started to feel... a tiny bit slow during the highest overhead periods. My E6300 at 3.5ghz really doesn't show ANY lag at ANYTHING :p
Click to expand...

Thank you for responding.

I am curious what is your FSB?

Like I said in my previous post, I should have asked a different question...

I should have asked more about the FSB being a bottleneck in heavily multitasking scenarios vs Direct Connect, and HTT.

Thanks
 
Ahh, well, you're right. the C2D is MUCH more bandwidth limited than the A64 platform. However, the clock for clock performance of the C2D is much higher. For example...

Both my X2 and my C2D were run at the same clockspeed for testing, and the Intel system was *much* more responsive. I was able to increase the number of live effects in Cubase by literally 100%.

The FSB sucks on the C2D, but the processor is so much more efficient, it makes up for it and then some.

My FSB right now is set to 500. That's 99mhz over the 1333mhz strap, so my memory bandwidth is quite high. If you dont know what the strap means... I suggest you google what this means as I'm quite clueless myself... All I know is that, at certain FSB ratios it divides down the speed that the memory and NB operate at... Sigh, still quite clueless.

Anyways, it all boils down to this: From what you've described, the "multiple tasks" dont seem to be really using that much memory bandwidth! Rendering a 3D scene, while exporting a PDF might create a memory shortage, but for the most part, your applications are reading a bit of data from the HD, doing lots of small calculations etc. Nothing that will stress how your memory is accessed.
 
EQTakeOffense said:
http://www.anandtech.com/memory/showdoc.aspx?i=2800&p=1

The article above basically shows that even though Conroe has (generally speaking) crappy bandwith, it still outpreforms the NetBurst and K8 in just about every application clock for clock.
Click to expand...

How do you define crappy ?

Is 5GBs vs. 7GBs considered "crappy" ?

800MBs for a P3 is crappy compared with 5 GBs for a Conroe...but calling Conroe's BW crappy is a little too much IMO.
 
Arcygenical said:
Ahh, well, you're right. the C2D is MUCH more bandwidth limited than the A64 platform. However, the clock for clock performance of the C2D is much higher. For example...
Click to expand...

As Jeremy Clarkson says : "utter rubish".

Adding GBs from all over the place is a fun exercise , but has little effect in desktops.


The FSB sucks on the C2D, but the processor is so much more efficient, it makes up for it and then some.
Click to expand...

That is a fallacy , no matter how powerfull an elephant is , drinking through a straw won't help .
In other words , the BW available isnt' great but good enough
 
Arcygenical said:
...Anyways, it all boils down to this: From what you've described, the "multiple tasks" dont seem to be really using that much memory bandwidth! Rendering a 3D scene, while exporting a PDF might create a memory shortage, but for the most part, your applications are reading a bit of data from the HD, doing lots of small calculations etc. Nothing that will stress how your memory is accessed.
Click to expand...

Thank you for the explanation, very helpful.:)
 
savantu said:
...In other words , the BW available isnt' great but good enough
Click to expand...

So whats your opinion on the FSB being a bottleneck in heavily multitasking scenarios vs Direct Connect, and HTT?

Will an AMD platform be better?

For everyone's help I want to thank you.

PS: Has anyone seen benchmarks of Heavy Multithreaded use, I emphasize heavy, not one of these benchmarks with AVG and Office

Thanks again guys/girls.
 
enumae said:
So whats your opinion on the FSB being a bottleneck in heavily multitasking scenarios vs Direct Connect, and HTT?
Click to expand...

No.FSB is a bottleneck when multiple CPUs share it and extremly heavy I/O is involved.Which means servers .

For desktops , the FSB is good enough that Conroe beats Direct Connect+IMC by 20% on average.


Will an AMD platform be better?
Click to expand...

No.
 
EQTakeOffense said:
I define crappy as, I would much rather have HTT then FSB when multiple cores come into the picture.
Click to expand...

So you'd rather have something that sounds great instead something that performs great? Sweeeeeet!
 
enumae said:
Thank you very much. :)
Click to expand...

I worked with .pdfs as well as different photo-CHOP stuff. More times than not, the bottleneck is caused by I/O more than the processor. A lot of RAM, Fast and Multiple Dives will help you.

Please don't believe stuff AMD guys say about the FSB. I know Duby229's hate for it is misplaced and misguided.
 
Bandwidth is meaningless. Think of AMD and Intel as having switched spaces: Intel does more work with it's limited bandwidth, AMD does less work per cycle so it makes up for it with double-digit bandwidth. :) Don't worry about it.

EQTakeOffense said:
How is FSB better then HTT?
Click to expand...
It's not, and it doesn't have to be. See above for details.
 
You must log in or register to reply here.
Back
Top