SRL Archive - Discussions & Debates

We all know Intel owns AMD in benchmarks, it's pretty much a fact.
I DON'T want this thread turning into an overall free for all AMD Vs Intel thread!

I'm wondering when you can actually notice/tell the difference between AMD and Intel when using a computer!

What kind of programs, applications, uses, etc would you be able to actually know/feel a difference?

Links to articles, proof, stats, etc is greatly appreciated!
I don't want just straight up benchmarks for processors being pushed to their limit and being better than others. I want to see time/FPS/whatever else comparisons of 'practical' uses where someone would notice.

Typed on phone sorry if typos.

I am on an AMD and I notice a bit of lag. I'll be getting an SSD in 2 weeks and hopefully that will prove it's just the 5400RPM harddrive that is the bottleneck.
http://www.samsung.com/us/computer/laptops/NP535U3C-A01US

The only way you'll see the difference is when what you use require alot of CPU power.
Eg: Run 10+ smart at the same time.

But usualy you won't notice a difference if what you are doing isn't very CPU intensive.

intel chips currently have a lot better latency times on servers, from what I noticed we're able to get almost 25k requests/second more with a simple i3 (2core/4thread) processor vs any type of AMD processor (from the bulldozer architecture at least) Just simply because it seems the i3 processor handles the request and gets rid of it faster than the amd even recognizes it has a request, almost like tripping during a race.

In everyday, like gaming and stuff the graphics card takes almost all of the load and you really won't notice much of a difference maybe if everything is the same except processor you might get like 2 more frames with an intel? maybe.. (pretending both processors are close to equal in power, like two quad cores)

and as said something like 10+ smarts/simbas, which might as well be considered a benchmark because it's doing nothing different than eating as much resources as it can to run efficiently

I will assume there is no overlocking going on so:

In all truth despite what anyone on the planet says about Intel and/or AMD - you wont notice any difference in anything if you're using a quad-core Intel vs quad-core AMD at the same clock speeds assuming you're a general user but now, with running multiple applications and video rendering the Intel i series quad-cores will beat the AMD FX quad cores because Intel are using 2 threads per core which boosts performance when multi tasking considerably whereas AMD only have single threads on the cpu cores. (some i series cpus don't use multi threading though)

So in short the above posts are right about the 10 simba things, Unless you can max out a quad-core you wont be seeing much, if any difference between products.

As for feeling a 'proper' difference you would feel it more with an intel i5/i7 vs amd fx 4100/6100 while converting videos and other cpu intensive tasks that support multiple cores/threads properly.

Source: Sorry no links, but i'm running an AMD FX 6100 atm and really, there's no difference when I play on my Intel i7 rig until I start video converting/rendering. (both pcs are same ram and gpu and hdd)

Edit: The i7 is faster than my FX 6100 at video rendering/converting**

I know for most games, if that is all that is running, an i3 will beat out the FX-6100.

From what I gather Intel is better at Running more bots as they Hyper Thread as where AMD's don't.

AMD's bulldozer architecture for the 8core is almost identical to hyper threading to the point where people just call it a 4core/8threads

Bulldozer has horrid IPC, it's the worst any cpu has had in.. forever

that being said, they're doing the exact same thing hyperthreading does, they have a pool of resources per 2 "cores" (their cache for example) which is the same as hyperthreading.. AMD calls it "modules" each set of cores share a "module" a big problem with AMD cores, because they didn't want to call hyperthread, the cache is used between core 0,1 and then 2,3 etc.. the problem with this is the core 0 & 1 will be the first two cores used for running something.. guess what? That means all the other modules cache is sitting at 0% usage, vs it could use core 0 and 2 then it would have two separate cache pools

But still, IPC is really where bulldozer fails, and that's why they should have taken the hyper threading method instead of trying to be cool saying they had 8 cores..

So really, AMD can't hyper thread because they're already basically doing it.. if they were to hyper thread, 4 threads would share 1 cores worth of cache and whatever else(1 core, 2 threads - intel | 1 module, 2 cores - AMD)

Didn't read any of these, just googled so people don't have to:
http://en.wikipedia.org/wiki/Inter-process_communication
http://www.tomshardware.com/forum/316198-28-bulldozer-death-1000-cuts

typed that pretty fast, probably barely makes sense but I gotta go