AMD FX-60 Dual Core

[Allan]

Originally Posted by dsio

Very true. Cache is not the holy water that it is for Intels on an Athlon 64 machine, so any gain from the 1MB L2 is minimal at best.

Actually, my tests prove otherwise. In WinRAR a 4000+ single core with 1MB L2 is quite a bit faster than an X2 4600+ with only 512kB L2 on each core. Same goes for some games (Doom3, HL2). In HL2 the 4000+ scores an extra 10 fps, and 15 fps in Doom3.

In WinRAR it can compress 335 MB files 17 secs faster than the X2. (123 vs 140 seconds). Thats a 12% increase, I wouldn't exactly call that a "minimal gain".

[dsio]

Allan, that is not a fair comparison at all.

You make it sound like the only difference between the two is the 512K cache. That is not the case. Even with identical specs, for example comparing a 4000+ to a 4800+ on a single core task, the 4000+ has a significant performance lead.

Have a look here: http://www.anandtech.com/cpuchipsets...spx?i=2410&p=3

This isn't just anand, its everywhere. Dual core chips clock for clock, cache for cache are slower than single core chips when running single core tasks.

To make an accurate comparison, compare a 4000+ to a 3800+ single core.

There is simply no accurate comparison to be made in single core performance between a dual core and single core, when the dual core has such a large performance overhead.

When comparing the 4000+ to the 3800+, which are 2.4Ghz 1MB L2 and 2.4Ghz 512K L2 respectively, you will NOT find much of a difference at all.

Some select cases will give maybe 5%, but that is usually about as much as the double cache will give.

[Allan]

Originally Posted by dsio

When comparing the 4000+ to the 3800+, which are 2.4Ghz 1MB L2 and 2.4Ghz 512K L2 respectively, you will NOT find much of a difference at all. Some select cases will give maybe 5%, but that is usually about as much as the double cache will give.

Well, using the link you just stated, there is as much as:

11.7% difference in the Mozilla test.
8% difference in SysMark Communication test.
(Between 3800+ and 4000+)

I didn't say that it was always faster with double the cache, but that it in some cases gives way better than a minimal gain.

[dsio]

Ill take that point, but I don't consider 11% on a browser benchmark to be anything to write home about.

The fact is it is a much smaller factor on an Athlon 64 than it is on a Netburst chip, which is extremely cache dependent.

[madcatmk3]

AHHHH!!! I'm going crazy why can't there just be two options I'm now going crazy having to memorize the differance between all of these. I'm crying it's so hard, they keep malking new things. I'm starting to dispise my job.

[Áedán]

That's how these companies continue making money. Produce a new version, and let people buy the new version.

As far as despising jobs, I had one job that I managed a total of 6 weeks in. Never again I told myself! Fortunately, I have never had to do the same job since.

[madcatmk3]

I custommise or build computer for people so I have to know every piece of hardware and explain it to them. However I make a lot of money each time.
When explaining everything to somone it talkes them around 3 hours to get it done and I charge $20 and hour.
Still every day something new comes out and I have to be unbiast which is ahrd since I pre nvidia over ATI ect. thats why I find this place so useful since the fact thats all of you like on side or the other. But since I started things have changed quickly. When I first came here dual-core was in proto type stage, now it being replaced by 65nm dual core, it all so confusing.

[Áedán]

Yeah, I know what you mean. I work as a consultant, and we charge the client over $250 per hour in blocks of one day. We have to be vendor neutral, and know a great deal about various things. Fortunately, we work as a team, so no one of us knows everything, but we can get hold of someone who does.

[madcatmk3]

I work alone which is hard. But every once in a while I have a friend who helps me build a PC or something and we spilt the profit.

[Áedán]

Remember that a lot of the people here who keep up with developments do so on an informal basis, not as a job. If it is your job to know these things, then you should at least know of them, even if you do not know all the details.

There is no shame in saying "I don't know, but I can find out" if questions exceed what you DO know.

[Daniel ~]

Originally Posted by madcatmk3

I work alone which is hard. But every once in a while I have a friend who helps me build a PC or something and we spilt the profit.

Just keep in mind Aedan's insight.
You don't have to know everything today, you don't have to learn everything you don't know by tomorrow. The hardest part for guys like you and me...is keeping a calm steady mind. When we can do that, new information falls into place much more agreeably.

Stop and remind yourself that "They" didn't create all this yesterday! Millions have worked on this stuff for decades.

They say that the last man to know every thing about his world, the last world in which it was possible for even a brilliant mind to know everything human beings had learned, was over by the beginning of the Rennisaunce!
More than a thousand years ago... and we've learned a bit more since then! ":O}
Every thing after that came forward as a matter of team work and co-operation between men.

Last I heard Human knowledge is doubling every ten years.

A calm and collected mind will be the best tool you will ever have.":O}

[Gizmo]

Originally Posted by Daniel ~

A calm and collected mind will be the best tool you will ever have.":O}

Amen to that, Daniel.

The smartest man I ever knew (my Grandpa Vance) didn't know the answers to all of my questions (and freely admitted such), but he knew how to find the answers, if they were to be found.