I recently got a TR5 with a gig of Ram and was worried about the mobile 1.1ghz ULV processor, I do alot of video/Image editing and wasn’t sure if it would be up to the task.
In comparison I have a desktop P4 2.66 that was my main machine for doing work on.
Having painted the my picture for you, the TR5 drops a brown load on the P4, even if the TR is using 100% of the CPU putting a filter on a video, I can open other programs and check mail etc….. With the P4 the chip gets hot and the whole system locks up, I just have to wait…..
Is this the case with AMD chipsets? Or just a problem with P4’s, or am I a one off?
[quote author=“Godspeed”]I recently got a TR5 with a gig of Ram and was worried about the mobile 1.1ghz ULV processor, I do alot of video/Image editing and wasn’t sure if it would be up to the task.
In comparison I have a desktop P4 2.66 that was my main machine for doing work on.
Having painted the my picture for you, the TR5 drops a brown load on the P4, even if the TR is using 100% of the CPU putting a filter on a video, I can open other programs and check mail etc….. With the P4 the chip gets hot and the whole system locks up, I just have to wait…..
Is this the case with AMD chipsets? Or just a problem with P4’s, or am I a one off?
You have the wrong P4 for what you want to do. Your 2.66 runs at 400Mhz FSB, if you got the 2.4, 2.6, or 2.8 (all about the same price) with the 800Mhz FSB (Hyperthreaded) you could do several intense multitasks at the same time because the PC thinks you have 2 Different CPUs. It’s a night and day difference.
It depends on what you’re comparing…yes, a Pentium M 1.1GHz is no match for a 2.6Ghz Pentium 4. However, that doesn’t mean a Pentium M is a slouch. A newer Dothan-based Pentium at 2Ghz runs almost neck and neck with a 3.2GHz Pentium 4 (with it’s 800Mhz FSB).
A 2.6 and 2.66 are 2 vastly different CPUs. A 2.4 HT will step all over a 2.66…
When I made my comment on getting a HT CPU I wasn’t suggesting the TR was faster in any way, what I was saying was when the CPU is at 100% a HT CPU would be great when trying to open another app (which is what the post was asking about).
[quote author=“tifosiv122”]When I made my comment on getting a HT CPU I wasn’t suggesting the TR was faster in any way, what I was saying was when the CPU is at 100% a HT CPU would be great when trying to open another app (which is what the post was asking about).
I wouldn’t agree with that at all. HT is a nice hardware trick, but it’s not like you’ve added a new CPU. It just uses spare CPU cycles to emulate another proc. I’ve seen dual HT-capable Xeons (4 CPUs to the OS) brought to their knees due to a single process.
Oh, i wasn’t directing…just making a common statement that it’s hard to compare CPUs simply based on CPU speed. The architecture has a lot to do with it and what you’re running. Like Erik says, HT when enabled properly and using HT-enabled software can make things smoother.
However, it doesn’t necessarily mean it will be “faster”. A P4 2.6Ghz non-HT processor is faster than a P4 2.4Ghz HT processor in raw CPU performance…it’s just not as “smooth” as the HT system when under high loads. Of course, it requires the right conditions as HT has been shown to slowdown a system depending on the application. HT is not as effective as real SMP systems which are “creamy smooth”...or at least much more smooth even when under which loads. And simply adding a second CPU doesn’t double your performance. There’s always some amount of overhead.
Once we get real dual core chips and full NCQ enabled SATA-3 drives on our desktops then you will really know what “creamy smoothness” is…
And yes…that’s a technical term…CREAMY SMOOTHNESS!
[quote author=“gr00vy0ne”]Once we get real dual core chips and full NCQ enabled SATA-3 drives on our desktops then you will really know what “creamy smoothness” is…
And yes…that’s a technical term…CREAMY SMOOTHNESS!
Not if Microsoft has anything to say about it! By that time, we’ll have a fully DirectX-powered desktop full of dynamically-generated shadows and shader effects, an SQL Server-powered filesystem, 8384-bit encryption to get to your bank site (and enough spyware to fill your current hard drive).
[quote author=“Drachen”][quote author=“gr00vy0ne”]Once we get real dual core chips and full NCQ enabled SATA-3 drives on our desktops then you will really know what “creamy smoothness” is…
And yes…that’s a technical term…CREAMY SMOOTHNESS!
Not if Microsoft has anything to say about it! By that time, we’ll have a fully DirectX-powered desktop full of dynamically-generated shadows and shader effects, an SQL Server-powered filesystem, 8384-bit encryption to get to your bank site (and enough spyware to fill your current hard drive).
Yeah, so by then…we’ll need dual, dual-core chips. And our file-subsystems will be all RAID-enabled and we’ll probably need to go back to those caching controllers. Also, you have to figure at some point that video cards will go dual-core as well since those suckers are more complicated than these mundane CPU designs. You can see Alienware doing that already with their proprietary dual video card design (it’s like the updated SLI of the new millenium). :xeno:
So where do AMD CPU’s sit in the hierarchy, do they have anything that multitasks like a P4 with HT? If you had a choice what would be your CPU for a desktop…. for games/multitasking/encoding etc…
Next to the Pentium M CPUs, the AMD Athlon 64 processors have the highest performance per clock ratings. They are affordable and high performing parts. They also consume a lot less power than Pentium 4 chips so they naturally run cooler. Also, because the memory controller is integrated into the CPU, they have high memory bandwidth performance and low latency.
Of course, both AMD and Intel have their strength and weaknesses. In most general areas for the desktop, the AMD chips are faster than Intel’s at the same price points. The one area Intel P4 chips excel is in the multimedia stuff which are coded to take advantage of Intel’s HT and SSE3 extensions. These programs are few and far between so unless you know your program REALLY takes advantage of it then the AMD is a better bet right now.
Also, Intel is changing their entire socket structure and board configurations so you’re more likely to be stuck with obsolescense if you buy now.
AMD’s top-of-the line and Intel’s top-of-the line are very similar in performance. It REALLY depends on exactly what you are doing, and even then, much of the real performance you get will be determined by things other than the CPU, like the memory, mass storage, video card, driver versions, operating system, etc… To answer your basic questions:
Games - Athlon FX (socket 939)
Encoding - Pentium 4EE (better SSEx support)
Multitasking - (assuming you mean basic Windows use and Web surfing) whatever you want. Bang-for-buck, you should look at the Athlon 64 series.
Based strictly on hardware, ATi. I’m really feeling burnt out on ATi’s drivers lately. I stayed with the 4.2s for a while because the 4.4 and 4.5 screwed with desktop transparency apps. When the 4.6 drivers came out, Far Cry became decently playable with lots of effects on, but KOTOR took a dive. After finishing my round of Star Wars, I’ve so far stuck with the 4.6 drivers. Pain in the ass. Once I can find a dual DVI GeForce 6800 Ultra that’s actually shipping, I’ll probably dump ATi. Image quality seems much better on ATi at the moment, so the jury is still out at this point.
[quote author=“Anonymous”]Regarding the encoding (at least of DiVX), the P4 is currently still faster, but this is supposedly one area in which 64bit computing will make a big difference.
Depending on the encoding, yes. 64-bit registers will allow a CPU to store 128-bit precision floating point numbers in single FP registers and 64-bit integers in single registers, but that won’t do a damn thing unless you use those sizes. In theory, there would even be a slight penalty due to the fact that the size of the instructions and data is now 64-bit. All pointers are twice as large too, so you’re also using more memory. One of the biggest benefits to x86-64 is a slew additional registers to use. SSE instructions use additional floating point registers that didn’t exist in previous versions of x86, so compilers could use the floating point registers more efficiently.
That one depends. When you compare the two top models, the X800 is faster and has nicer AA. Furthermore, it consumes much less power than an 6800. But the 6800 of course supports SM3.0, although it has still to be seen of how much importance that will be.
On the encoding issue, I’m sure I read somewhere that x86-64 would bring substantial advantages, but I can’t seem to find any reputable sources, except some claims on a message board. One thing is that AnandTech tested a 32bit encoder on 64bit windows, and he already saw an increase of 13-15%. So there’s definitely more room for improvement.
GeForce/Radeon: Very true. All things being equal I’d go wih the X800, but they’re not. Dual DVI and AGP are both very important to me, and ATi and it’s licensees don’t seem to be delivering that combination. The driver issue is secondary, but still important.
Encoding: Not sure why that program would be faster in 64-bit mode. The games results seem to bear out what I said above.
Yeah, so by then…we’ll need dual, dual-core chips. And our file-subsystems will be all RAID-enabled and we’ll probably need to go back to those caching controllers. Also, you have to figure at some point that video cards will go dual-core as well since those suckers are more complicated than these mundane CPU designs. You can see Alienware doing that already with their proprietary dual video card design (it’s like the updated SLI of the new millenium). :xeno:
