• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

One whole core just for a game.

Dutch Guy said:
If you have a optimised system with no unneccesary services and tray programs you are right, but with a lot of crap running a dual core system (might be) is faster than a single core system.

So basically, you're saying dualcore is for lazy and/or stupid people?

;)

:p

j/k (in case you couldn't tell!)
 
Dutch Guy said:
Good point, AMD and Intel can make it a slogan:

Now just for the lazy, messy and uninterested people out there, a CPU that is just as fast as normal no matter how much rubbish you have running :D

Language used to describe people of which I presume do not view these forums too often ;)
 
Dutch Guy, jumping to conclusions here, but have you ever used a Dual Core.
I found that even on a clean install my X2 3800 was significantly quicker around windows than my Winnie 3000. (this was back when I only had a gig of RAM)
 
PinkFloyd said:
Dutch Guy, jumping to conclusions here, but have you ever used a Dual Core.
I found that even on a clean install my X2 3800 was significantly quicker around windows than my Winnie 3000. (this was back when I only had a gig of RAM)
As I said the only thing I remembered was seeing some reviews in the beginning, I have only recently upgraded to a Skt 939 setup.

Obviously a dual core CPU is faster at Windows stuff, it's just that games do not benefit enough from two cores which is a shame.
 
You must have had one of them days when your playing online then your anti-virus kicks in you get a big drop in Frames. I'm guessing with a Dualcore you won't get effeced.
I always end up stopping the scan so I can carry on doing what I was doing so I end up never letting my virus protection do its job.
 
I know what he's getting at but I think the main problem is that even if you have 2 cores its not like having 2 systems in one. For example they still share the same memory.... and im sure you lose a fair amount of speed because of that.
 
Synergy said:
I know what he's getting at but I think the main problem is that even if you have 2 cores its not like having 2 systems in one. For example they still share the same memory.... and im sure you lose a fair amount of speed because of that.
Well you can do (lose a bit of speed) but luckily I believe each core on the X2 has it's own 64bit memory controller (please correct me if I'm wrong) a reason why the X2 bandwidth benchmarks are always off from single cores at the same speed. (64 bit is single channel, 128bit is dual channel). Thus, whilst they share the same memory, if you have 2GB, then they aren't so badly affected as they don't share a memory controller like the Intel dual cores.

The slowdown comes from the fact that they use the same bus meaning the cores have to queue their instructions but the memory controllers act independently.
 
Last edited:
PinkFloyd said:
Dutch Guy, jumping to conclusions here, but have you ever used a Dual Core.
I found that even on a clean install my X2 3800 was significantly quicker around windows than my Winnie 3000. (this was back when I only had a gig of RAM)

Dual core @ 2Ghz + 2GB RAM vs Single core @ 1.8Ghz and 1MB RAM

Hardly a fair comparison. I suspect a lot of the speed was from the 2GB RAM, hence less disk swapping.
 
smids said:
Well you can do (lose a bit of speed) but luckily I believe each core on the X2 has it's own 64bit memory controller (please correct me if I'm wrong) a reason why the X2 bandwidth benchmarks are always off from single cores at the same speed. (64 bit is single channel, 128bit is dual channel). Thus, whilst they share the same memory, if you have 2GB, then they aren't so badly affected as they don't share a memory controller like the Intel dual cores.

The slowdown comes from the fact that they use the same bus meaning the cores have to queue their instructions but the memory controllers act independently.
Sorry, I wasn't being specific (as usual :P) - yes this is exactly what I mean, them using the same bus does mean a hit to the system and of course you do need enough memory to go round. There is of course an added slowdown attributed to multiple programs accessing the same hard drive at the same time. This may or may not be likely to happen depending on your setup and what you are running.
 
Synergy said:
There is of course an added slowdown attributed to multiple programs accessing the same hard drive at the same time. This may or may not be likely to happen depending on your setup and what you are running.
No there isn't. NTFS uses an elevator queue in scheduling disk head movements. Also, since the NT kernel can dispatch hardware interrupts on both CPUs, one CPU can be used for the I/O dispatch whilst the other continues processing other threads. If the software using non-blocking asynchronous I/O then there is actually a large gain to be had. A lot of Windows software is actually designed to use this.

Jimbo Mahoney said:
Dual core @ 2Ghz + 2GB RAM vs Single core @ 1.8Ghz and 1MB RAM

Hardly a fair comparison. I suspect a lot of the speed was from the 2GB RAM, hence less disk swapping.
While that is a valid point it doesn't really matter anyway. It's a known fact that an SMP system is faster around Windows than a uniprocessor one. Back in the late 80's and early 90's when the NT kernel was under intense development a lot of work went into ensuring it would scale well on SMP hardware. As a result, the entire NT kernel as well as third party device drivers can execute across multiple cores. More traditional operating systems can only execute the kernel mode on just one processor in the system at a time. This fact alone places NT at the relative forefront when it comes to scalability on SMP hardware. The Explorer shell as well is highly multi-threaded (just take a look at how many threads it has open in Task Manager) which goes a long way on SMP hardware. In fact, Windows software as a whole tend to be very highly multi threaded in comparison to other platforms. Again, just take a look at Task Manager...
threads.gif
 
Last edited:
NathanE said:
No there isn't. NTFS uses an elevator queue in scheduling disk head movements. Also, since the NT kernel can dispatch hardware interrupts on both CPUs, one CPU can be used for the I/O dispatch whilst the other continues processing other threads. If the software using non-blocking asynchronous I/O then there is actually a large gain to be had. A lot of Windows software is actually designed to use this.

So you're saying there is no slowdown due to two processes PHYSICALLY accessing the disk at the same time? Lets say for instance, you are playing a game on one core. The other core is encoding a movie stream. Both write to the same HDD, there is bound to be some hitching that would not normall be encountered when the game needs to cache data.
 
Yes that's what I'm saying. I am also saying that, if anything, all forms of I/O (including disk access) will potentially be performed faster on an SMP system.

There is no hitching because disk I/O is inherently an asynchronous process and any (even the crappiest of game engines) are written with that mind. It is such a fundamental concept in computing. Software does not just say "I want to read blah.dat and I want it NOW.", it says something more like "I want to read blah.dat and have the results on my desk in less than a second."

What are you saying? That if you were to encode a movie and play a game on a uniprocessor system that it would complete faster than on an SMP system? :p
 
NathanE and Synergy, you have been arguing the same thing - obviously with Nathan talking a language no-one understands :p (j/k, you do a good job at simplifying things) and Synergy the same thing.

[Telepathic goggles on]
All Synergy was saying is that running a DC (SMP) system does not mean it is like 2 systems in one. There is some slowdown as opposed to using a dual processor. He was merely saying don't expect too much from it, although there will be a vast improvement on single core.
[/Telepathic goggles off]
 
smids said:
NathanE and Synergy, you have been arguing the same thing - obviously with Nathan talking a language no-one understands :p (j/k, you do a good job at simplifying things) and Synergy the same thing.

[Telepathic goggles on]
All Synergy was saying is that running a DC (SMP) system does not mean it is like 2 systems in one. There is some slowdown as opposed to using a dual processor. He was merely saying don't expect too much from it, although there will be a vast improvement on single core.
[/Telepathic goggles off]

Thankyou! :D Exactly what I was getting at.
 
smids said:
[Telepathic goggles on]
All Synergy was saying is that running a DC (SMP) system does not mean it is like 2 systems in one. There is some slowdown as opposed to using a dual processor. He was merely saying don't expect too much from it, although there will be a vast improvement on single core.
[/Telepathic goggles off]
A dual CPU system will have the same disadvantage of having to share the memory bandwith :confused:
 
Back
Top Bottom