Computer Build

I was comparing the price for that 6 core intel I7 to the AMD 6 core 1100. The bench mark isn't that much different from the AMD to the Intel, what .4 better on the bench mark. Also you are comparing the I5 and that is only a 4 core. I'm not trying to argue with you. I'm just saying for the money that was posted on the earlier posts I would go with the AMD.

I just felt you were giving some misleading info when you say there was $800 difference in comparable processors. That my be true, if you're comparing only hex core CPUs between AMD and Intel. But I was comparing the closest comparable performance to price ratio with the chosen proc you own and run, which I feel offers better advice to members and guests seeking information on this forum. This is not to win an argument for me, just to offer info that is best suitable for people possibly reading this thread.

 

It's the end performance to price ratio that matters, not cores or threads or cache or floating point algorithms or whatever else you want to compare side by side.

 

And for the record I do like AMD procs. Intel is still my #1 choice overall but that's merely my opinion. I just believe AMD has always been the bargain proc, unless you count 4-5 years ago when they were whipping Intel's butt before the Core came into the lineup.

And yes for the record, the O.P. doesn't need to spend a grand for that proc. Core i7 SandyBridge is all he needs at best. I don't know why I didn't offer that piece of info in the begging now that I think about it... lol

And yes for the record, the O.P. doesn't need to spend a grand for that proc. Core i7 SandyBridge is all he needs at best. I don't know why I didn't offer that piece of info in the begging now that I think about it... lol

Yep! And even compared to AMD chips this one's not a bad buy: http://www.cpubenchmark.net/high_end_cpus.html

 

The closest AMD Phenom II x6 1100T for $214 is still lower ranked than even the i5-2500 @ $210, and for a mere $84 extra you could go with the i7-2600 @ $298. It's when you want to go with the "reccomended" Xeon chips where prices start creeping into the $1000 bracket (especially for those surpassing the i7-gen 2 chips). AMD's CPU's seem to be falling by the wayside, hopefully they're working on some new ones for the future - but they're possibly swamped with trying to get their new baby (the GPU's) to surpass nVidia's.

 

Talking of which: http://www.videocardbenchmark.net/high_end_gpus.html

It's not necessarily the most expensive which gives the best performance. And it's not needed to have the absolute highest GPU out there if your 3D work isn't EXTREMELY complex.

 

As for multi-core: ACad uses it only in redering, the regening is a near-farce since that happens very seldom in new ACads. So if you don't render, then having a platoon on cores won't help much. That said, having at least 2 cores makes your PC more responsive - e.g. say ACad is working on something which may take several minutes. If you only had one core, it would feel as if the entire PC is dead. While with a second core, you could still go do some emails. The more cores the more such things can happen concurrently.

 

The main thing is to get the GHz as high as possible per core. That's why the mobile version CPU's aren't too great for CAD work. Their GHz rarely go above 2.2, while these i7's (and the Xeons & Phenoms) go above 3 mostly. This isn't the only factor - some CPU's have faster bus speeds so data is sent & received quicker from RAM (which might help speed up the workload as well), so it's not a direct correlation between higher GHz and faster computing. That's mainly why a P4 @ 3.2GHz is outperformed greatly by a Core2 @ 2.3GHz - the bus speeds are increased from 533MHz to 1333MHz between those 2 (not evewn considering 64/32bit), while a i7-2600 @ 3.4GHz can go up to 2133MHz bus (depending on your RAM and some MB settings). And then the caches also come into play, usually a larger cache means less IO between the CPU and RAM. And lastly some of the optimizations could also be used, though these usually have a lot more effect when working with video - while AMD's have something called 3dNow since a while back.

 

But yes, get the best CPU you can afford. I'd also state go with larger (and possibly faster, e.g. 1600MHz instead of 1066MHz) RAM instead of a HUGE GPU - do get a "capable" GPU though. And then having 2 HDDs helps quite a bit (note NOT 2 partitions on the same HDD), and would help even better if the one containing your OS / Programs / swapfile(s) is a high-speed HDD / SSD (even if it's small). E.g. the newest SSD's are only now starting to reach the Sata II interface speeds of 3GBits/s - meaning throughput gets to around 380MBytes/sec (at best). In which case you might start looking for Sata III discs at 6GBits/s (or "perhaps" more) in the future.

Talking of which: http://www.videocardbenchmark.net/high_end_gpus.html

It's not necessarily the most expensive which gives the best performance. And it's not needed to have the absolute highest GPU out there if your 3D work isn't EXTREMELY complex.

There is no way that test is accurate. The workstation level graphics cards utilize different OpenGL extensions that are not available on the gaming cards. That test said "everyday computing", even in the advanced testing. Someone needs to test gaming cards against workstation cards utilizing real world 3D modeling, video processing, and rending applications other than games.

Tanner, remember that post I made with the graphics card test?

 

http://www.cadtutor.net/forum/showthread.php?51452-Graphic-card-comparrison-from-CGArchitect.com

 

http://www.cgarchitect.com/news/Reviews/Review076_1.asp

Nice. Never saw that post you made. Looks like gaming cards win in that real world test.

 

That should be a sticky.

Yeah, it's when you run shademode that the Quadros come through. Now if some one is not doing 3D then a game card would be the best option and trade-off.

I still like my Quadro though

And I still like mine too. I do run shademode quite a bit, since I'm in deep into 3D BIM design on a daily basis.

I haven't tested it again with the new Quadros, but previously (about 3 years ago) we had some FX550's and on my Laptop I had a Radeon X1600. ACad performed only slightly better on extremely complex models with shademode on for the Quadro. But where we saw that the extra expense was wasted was when running Revit with its shademode / shadows turned on - here there was literally no difference whatsoever between the 2 cards (and that was on a 120MB model of a 12 floor hotel with lots of corbelling detailing) - using Revit 2008. Not to mention the quadros tended to produce artefacts all the time, where the Radeon never did (and we were using the latest drivers at the time).

 

Perhaps with the new drivers as stated in that test there may be a bit more difference noted. Though that test says nothing about Revit, only Sketchup, which I assume would be much similar - i.e. not much difference! Anyone have experience on how Revit 2011/12 performs between the pro / consumer cards?

I need to install my 2012 Arch suite (Revit 2012, AutoCAD Arch 2012 and AutoCAD 2012) and see how things run with my new machine.