Jump to content
bioduken

AMD vs. Intel Architectural Debate

Recommended Posts

He might be asking for overclocked scores from the Q6600.

 

FYI that is a Go Q6600 that was tested. :)

 

AMD has much better I/O performance. Faster GPU cards, higher resolutions or what things that will increase the communication between the processor and in this case the GPU will need more I/O.

The FSB running at 333 MHz has a theoretical bandwidth of ~10 GB/s, overclock to 400 MHz and that bandwidth is increased to ~12 GB/s.

Hypertransport on AMD has about ~20 GB/s and it only needs to handle I/O.

The FSB on the intel needs to handle both memory traffic and I/O. This is also the reason why you need to overcklock Intel to get as much performance as you can. You don

Share this post


Link to post
Share on other sites
AMD has much better I/O performance. Faster GPU cards, higher resolutions or what things that will increase the communication between the processor and in this case the GPU will need more I/O.

The FSB running at 333 MHz has a theoretical bandwidth of ~10 GB/s, overclock to 400 MHz and that bandwidth is increased to ~12 GB/s.

Hypertransport on AMD has about ~20 GB/s and it only needs to handle I/O.

The FSB on the intel needs to handle both memory traffic and I/O. This is also the reason why you need to overcklock Intel to get as much performance as you can. You don

Share this post


Link to post
Share on other sites
Somehow I predict some interesting responses to your last two statements. Would you care to enlighten us further on the basis for your claims?

 

Here is one: The need for an IMC and why the FSB is dead

 

Intel is very fast as long as memory is in the cache and threads don’t communicate (quads). But when it needs to communicate with external hardware (memory, GPU, manage cache coherency for quads) it needs to go through the FSB. If the FSB is occupied with other work then there are additional delays. All these new fast video cards can handle a lot of I/O and there will be much more traffic in the FSB. Games that scale to more threads they need to synchronize memory (spinlocks etc). On C2D this isn’t a problem more than it just can run two threads simultaneously. Intels Quads are just two duals glued together and they communicate using the FSB. Intel Quads also have a little more latency when it needs to communicate with external hardware.

When the FSB is choked it isn’t that effective. So when performance matter it is vital that the FSB is as fast as possible otherwise other hardware just sits and wait for data to arrive.

When you increase the speed of the FSB the computer feels faster because it responds faster. Compare this to AMD which has very fast responses. When the computer is booting up, starting new applications etc, the cache is not that important. Applications often has initialization code that is used to init the applications and isn’t used after that. There is a hype that one processor needs to be as fast as possible. All processors today are VERY fast, it isn’t the processor that is the main bottleneck for speed (except for some very special applications, mainly applications that do heavy calculation on small data chunks)

Running many applications or using memory intensive applications is also something that needs performance to and from the processor and other hardware (in that case it is the memory). If cache is trashed with data from a lot of applications or a lot of threads that use data then the cache isn’t that effective anymore. Intel has a very fast and big L2 cache but that cache is simple. It can’t synchronize four cores. The AMD L3 cache is more complicated, it can handle cache coherency and is used with the memory controller on the processor. AMD has used a lot of techniques to decrease latency on memory. In threaded environments or other memory intensive tasks this is good, but compared to raw calculation power not using memory it isn’t that effective.

 

More links:

http://en.wikipedia.org/wiki/Front_side_bus

http://en.wikipedia.org/wiki/Cache_coherency (movie: http://www.intel.com/technology/quickpath/demo/demo.htm )

http://en.wikipedia.org/wiki/Hypertransport

Edited by bioduken

Share this post


Link to post
Share on other sites

The really bad thing is AMD does very well with some of their stuff but as for having anything comparable to Intel they just don't have it.

 

If you compare AMD's slowest Quad to Intel's its a bloodbath and that is without having to overclock the Intel as you say.

 

If you take AMD's fastest CPU they make and put it against Intels, AMD is not even in the fight.

Share this post


Link to post
Share on other sites
The really bad thing is AMD does very well with some of their stuff but as for having anything comparable to Intel they just don't have it.

 

If you compare AMD's slowest Quad to Intel's its a bloodbath and that is without having to overclock the Intel as you say.

 

If you take AMD's fastest CPU they make and put it against Intels, AMD is not even in the fight.

 

AMD is much better when it comes to run demanding applications or applications that are threaded. It isn’t developed to run single threaded applications like Intel main target is. You need to understand this when these two processors are compared. It is also optimized for 64 bit performance.

It is much simpler to build a processor running single threaded applications then build a processor that will scale for multithreaded solutions.

Start a few applications that is working and measure performance, then you will see different scores in the battle between Intel and AMD.

Edited by bioduken

Share this post


Link to post
Share on other sites

Your theories are all well and good except for one thing... they're all rubbish and you're wrong lol

 

Feel free to post up any benchmark of an AMD quad core and I'll smash it to pieces with my Q6600...

 

If you want any sort of performance war and you have an AMD, go buy yourself a box of tissues because you're going to lose...

Share this post


Link to post
Share on other sites
Your theories are all well and good except for one thing... they're all rubbish and you're wrong lol

 

Feel free to post up any benchmark of an AMD quad core and I'll smash it to pieces with my Q6600...

 

If you want any sort of performance war and you have an AMD, go buy yourself a box of tissues because you're going to lose...

 

What is your explanation that AMD with a lower clock and much less L2 cache wins over one faster intel when the resolution goes up and/or the game is heavy on the graphics?

Edited by bioduken

Share this post


Link to post
Share on other sites

What CPU's are you comparing?

 

If you want to challange me any AMD you have against any Intel I have I will go head to head with you :D

Share this post


Link to post
Share on other sites
What CPU's are you comparing?

 

If you want to challange me any AMD you have against any Intel I have I will go head to head with you :D

is it ok to challange you on I/O performance? are you good at maths?

 

On I/O I can challenge you on a slow 9500 and you can choose any Intel you like, you won't win any way

Share this post


Link to post
Share on other sites

You said AMD has much better I/O performance. Faster GPU cards, higher resolutions or what things that will increase the communication between the processor and in this case the GPU will need more I/O.

 

Big deal.......

 

You said its faster based on that give me any AMD CPU and I will beat you to a pulp in gaming and just about everything else I throw out.

 

You might get me in memory scores but thats about it.

 

Anybody that buys an AMD to say that I have memory scores that are higher then you but gets rocked in everything else is not with it.

 

Anyone who knows anything there is no challange from AMD when it comes to CPU performance, now as for Video Cards that just changed. :)

Share this post


Link to post
Share on other sites

Both have their strong suites. Unfortunately for the kinds of apps and in most games the Intel will win out Period. Try video encoding. Yes the Phenom coupled with a Crossfire X setup on the Spider Platform is fast at video encoding but it is done by offloading the bulk of the work from the CPU and letting the GPU do the work. Clock for Clock the Intel Quad core is faster than a phenom in my book. I have no hardware allegiance. I will buy what performs best with what I do.

Share this post


Link to post
Share on other sites

Heres a little math for you, superpi 1m while all 4 cores are loaded 100% crunching with BOINC. If you can beat that with your AMD show me then I will stop BOINC and rerun it...

 

20080630194411jq0.png

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.


×
×
  • Create New...