Explain something

Page 2 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.

What would you do for a Serious Gameing rig?

  • AMD fx-60

    Votes: 14 35.9%
  • AMD Opteron 165 OC to 2.7

    Votes: 25 64.1%

  • Total voters
    39

TabrisDarkPeace

Distinguished
Jan 11, 2006
1,378
0
19,280
All decent dual-core x86/x64 processors support x64 via EM64T or AMD64 anyway.

The early Core Duo might not - the Pentium M dual core one.

Other than that nothing comes to mind as a dual-core x86/x64 processing lacking x64 support.

Can anyone else think of any others ?
 

luminaris

Distinguished
Dec 20, 2005
1,361
0
19,280
I think alot of it has to do with the fact that people want the most bang for their money without having to overclock and, they think it gives them bragging rights. Most people have no clue as what an Opteron is and would never think to use it as a regular workstation vs. a server.

If only everybody read this forum, they would certainly know the differences. On the other hand, if everybody that currently owns an fx60 owned an opty instead, they're might not be enough to go around.
 

hergieburbur

Distinguished
Dec 19, 2005
1,907
0
19,780
From my years of experience, here is my take:

FX-60 is geared toward performance gurus, enthusiasts, and people who have more money than know-how. As such, the core rating is determined by a number close to its actual maximum. The processor isn't really designed to have a lod of OC headroom.

Opterons are server chips, and are designed for stability under constant load. They are more than capable of handling very intense processes such as games (contrary to what some here have claimed). However, they are generally rated fairly conservatively in order to insure that stability under load. This leaves a lot of headroom for OC while still ensuring stability. You see this a lot with servers. Intel and Sun do the same thing to one extent or another.

To clear up some confusion, and I know some of this has been said already, but there is NO Instruction set differences, No architecture difference, and No performance difference between the socket 939 Opteron 1xx series and the Athlon 64/FX series other than they cache and the unlocked multiplier. The pricing is designed to be attractive to businesses, and those that keep up to date will remeber that AMD never intended to Opterons to get into the retail space, and has tried to limit sales to home users in order to have more available for business use.
 
Joefriday what u don't understand is that Eventhough ur looknig underthe hood and its Motor. You have'nt looked at its Shell. Its a GAMEING CPU and OPTY is a SERVER CPU. Different MOBO's too. Thats like Comparing a Dodge VIPER and a DODGE RAM clocked at the same Horse power. They might have same SPECKS and Engline But their a Totally Different Machine Ment for Different things AMD is obviously smarter than a Couple of Turds like Us who just RANT about their Processers.
Your ignorance is amazing; there is no real difference between an Athlon and an Opteron.
You rant about Opterons not being stable; last time I checked, servers are meant to be run 24/7.


AMD opteron is basically the same as the athlon 64, however, the opty has more hyper-transport links for a faster connection between the CPU cores and everything else plugged into the motherboard.
 

parlee

Distinguished
Dec 16, 2005
1,149
0
19,280
i think he means multipliers, which pretty much mean notihng if u up the memory speed u can achieve the same clock speeds as say a chip with 13 multiplyers with a chip that has 11 by simply upping the memory
 

hergieburbur

Distinguished
Dec 19, 2005
1,907
0
19,780
they both have 2000mhz ht... old socket 754 athlong 64's had 800 (1600mhz) ht. ur wrong, they are identical

Actually, that's not exactly right. It was determined buy the chipset that was used on the mobo, not be the socket that the chip used.
 

gOJDO

Distinguished
Mar 16, 2006
2,309
1
19,780
Although the CPU's you are trying to compare heve the same Toledo Core (and this applys to all cores and manufactures) The Opteron DOES NOT compare dirrectly to the Athlon 64, FX or X2 CPU's. The one thing that everyone seems to be missing here is the Instruction Set of the CPU's. Server based processors carry a different set of resident instructions then all destop processors. They are not designed for resident handling of many of the processes utililized in gaming. Yes, they will run most games quite well. However, the bulk of it's 'horsepower' sits idol while it parcels out opperations to the rest of the system in a game enviromrnt. While in the same enviroment, the desktop CPU's such as the FX handle this operations themselves.

1. What are the differences in the implemented instruction set of the Athlon64 FX-60 and the Opteron with the same (Toledo) core?
2. How do the Opteron parcels out operations to the rest of the system, what operations are that, what is the rest of the ystem and what is processing them?


After produced, chips are beeing tested and those who have passed the test are divided in quality classes depending on their ability to run stable at higher freqfency and lower voltage.
From the best cores, Opterons are made. They are suposed to process a lot of data, 24 hours a day. For stability, the Opterons are clocked less than their maximum at the current voltage and becouse of this they are so overclockable. For aditional stability and avoid of memory operation errors, the s940 Opterons are using Registered ECC buffered DDR, while the s939 are using unbuffered/ECC. Only the Opteron 2xx and 8xx have more than 1 HTT link enabled(each HTT link provides up to 8GB/s data transfer, the 2xx have 2 and teh 8xx have 3 links), while the 1xx and all other K8 processors have only 1.
From the next quality class, the AthlonFX are made. Games are CPU-power hungry apps and they are exploatating almost all the CPU resources, so stability again is an isue. The fully unlocked multiplier is an overclocking adventage over the rest K8 CPUs. Most of them are unlocked for lower multipliers or totaly locked(as the Semprons which are not supporting CNQ).
From this class are made the Turion64s also. They are power concerned, so stability at lower voltage is needed.
The remaining chips are for the Athlon64/64-X2 and the Sempron.

All the K8 with same cores on the same "the rest of the system"(mainboard, ram, etc.) and same clock are performing same.
 

TabrisDarkPeace

Distinguished
Jan 11, 2006
1,378
0
19,280
He is write, and that is why they can aggregate multiple dual-channel memory controllers performance. (eg: 6.4 GB/sec x 4 nodes = 25.6 GB/sec)

As you can see from:
http://www.amd.com/us-en/Processors/ProductInformation/0,,30_118_8796_9240,00.html
; they have more HyperTransport Links, but only on the Opteron 200 series and Opteron 800 series.

You can then view the presentations at:
http://multicore.amd.com
; for a better overall picture as to why they do this, and why they only do it on the 200 and 800 series Opterons.

Opterons 100 series on Socket 939 are explained here:
http://www.amd.com/us-en/Processors/ProductInformation/0,,30_118_8826_9384~85257,00.html
; As they (well the Socket 939 chipset platforms) do not support Registered ECC memory, only Unbuffered (ECC or non-ECC) memory.

Opteron 100 series on Socket 940 only really differ in the Registered (ECC is optional on both Reg and Unbuffered DIMMs btw) memory requirement, and thus total price, and maximum installable memory is higher using Registered sticks.

The entire Opteron 100 series (both s939 and s940) share the same HyperTransport setup as the current Athlon 64 series of processors do.