• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Opteron's big weakness

I understand completely about the co-workers knowing more about the hardware than the bosses do, in general, they deal with it more often, more experience by the virtue of job placement.

I'll explain my current mindset a bit, I am consulting with the National Guard for an instant on portable network. No problem. Piece of cake... Except for every damn recommendation we give is changed by some higher up that has to "upgrade" the requests. Instead of a C640 or D600, we get the D800, much larger and the only upgrades from the D600 is the damn LCD and video card.
Request a CDRW, we get modular Zip drives. Ask for a hard drive carrage (sp) so you can easily swap between workday and drill OS installs, we get modular bay drives which block access to floppy and CD drives.

Sorry for the rant... crappy day.

Anyways, this is the kind of management that has to be overcome for AMD to grow in large markets. I'm having a beer gentlemen, blood pressure is way too damn high...
 
I was going to PM you earlier Morley, but it was full. I have cooled down, so don't expect anything from me :)

Congratulations AMD. Now they can advertise something that maybe Intel doesn't have this month. I would imagin they could with in a few weeks with their resources :D

Vette
 
I agree with wench00 in that if AMD could licence hyperthreading from Intel the opterons would be an absolute killer server solution
 
Originally posted by 0ldman
I understand completely about the co-workers knowing more about the hardware than the bosses do, in general, they deal with it more often, more experience by the virtue of job placement.

I'll explain my current mindset a bit, I am consulting with the National Guard for an instant on portable network. No problem. Piece of cake... Except for every damn recommendation we give is changed by some higher up that has to "upgrade" the requests. Instead of a C640 or D600, we get the D800, much larger and the only upgrades from the D600 is the damn LCD and video card.
Request a CDRW, we get modular Zip drives. Ask for a hard drive carrage (sp) so you can easily swap between workday and drill OS installs, we get modular bay drives which block access to floppy and CD drives.

Sorry for the rant... crappy day.

Anyways, this is the kind of management that has to be overcome for AMD to grow in large markets. I'm having a beer gentlemen, blood pressure is way too damn high...
Click to expand...

At least your management didn't order 8MB AGP DVI-only S3 video cards for an Autocad lab with brand new 17" CRTs. ;)
 
Originally posted by Josh_B
At least your management didn't order 8MB AGP DVI-only S3 video cards for an Autocad lab with brand new 17" CRTs. ;)
Click to expand...
lol
not the same, but been there done that.
Ask for 20gig laptop hard drives, they find the cheapest 20gig drive out there... 3.5 inch hard drive...
Actually working on some small form factor Dell's for someone else, they order Quantum Bigfoots cuz they got a deal... drive no fit.
 
Originally posted by 0ldman
lol
not the same, but been there done that.
Ask for 20gig laptop hard drives, they find the cheapest 20gig drive out there... 3.5 inch hard drive...
Actually working on some small form factor Dell's for someone else, they order Quantum Bigfoots cuz they got a deal... drive no fit.
Click to expand...

That's funny.... but the management at my work spent months with the techs trying to figure out why Autocad was messing up the OpenGL rendering. Then I installed S3's latest 'driver', if you can call it that, and wow! magic! It worked!

1) Why did they order only an 8MB card?
2) What in the world compelled them to get the S3 card for Autocad?
3) What were they thinking buying a DVI card for a VGA-connected monitor?
 
I think trust in a brand name is relevant when there are hundreds of choices, but the mentallity behind it for when you only have two choices comes down to the fact that intel has X amount of money more than AMD does in the PR dept. where AMD were investing in innovative processor designs to keep the corporate and home secotrs alive with x86, intel put money into the PR dept and the itanium to get the corporates to change and move away from the x86 structure.

AMD did somthing intel didnt want to happen and that was keep the x86 alive for another 10 yrs.

my 2c ... ill shut up now.
 
Originally posted by @trapine
I agree with wench00 in that if AMD could licence hyperthreading from Intel the opterons would be an absolute killer server solution
Click to expand...

I would like to repeat what GodsMadClown has already said. AMD does not need hyperthreading. Hyperthreading is just a solution to the problem of having a pipeline that is too long in the first place.
 
Originally posted by DocFaustus
I would like to repeat what GodsMadClown has already said. AMD does not need hyperthreading. Hyperthreading is just a solution to the problem of having a pipeline that is too long in the first place.
Click to expand...

Have you heard of the next generation Dual-Core CPU's that AMD is going to be implementing soon? Two virtual cores aren't going to be able to do shit against a real integrated SMP chip with two sets of short, efficent pipelines, integrated memory controller and 64-bit instructions.

Hyperthreading will be another buzzword lost along the banks of technology come a year from now. AMD is looking to knock this one out of the park.

Matt.
 
Originally posted by enraged78
Have you heard of the next generation Dual-Core CPU's that AMD is going to be implementing soon? Two virtual cores aren't going to be able to do shit against a real integrated SMP chip with two sets of short, efficent pipelines, integrated memory controller and 64-bit instructions.

Hyperthreading will be another buzzword lost along the banks of technology come a year from now. AMD is looking to knock this one out of the park.

Matt.
Click to expand...

Yes, I have, but I heard about Intel's dual core Xeon first. Now that Nocoma has 64bit, then the dual core Xeon will too. Not to mention it will also have Hyperthreading.

Same goes with Itanium, sans HT.


I dont know how they do it, but whenever Intel does something, AMD has been planning it as well, and when AMD does something, Intel also has been planning it as well. I swear, there is probably some huge ub3r-corporation behind the two that owns both of them.
 
Great minds think alike. Intel and AMD are in the same world at the same time, with roughly the same technology availible. I hardly think it is suprising that they share some of the same engineering strategies. Multi-core is an old idea, and a VERY different matter than HT.
 
the reason they are so close is that they have cross licensing agreements for the x86 architecture, that is why mmx and sse and sse2 are in amd64 chips and x86-64 is in nocona, they are obligated to give anything new they add to the x86 architecture to the other company for free.
 
Supposedly AMD is getting SSE3 in the next revision or the one after that, continuing the tradition...
 
Well considering ATI and nVidia get together to agree on a general time-frame for the release of new vid cards, the relationship between Intel and AMD is not that surprising.
 
Originally posted by xonik
Supposedly AMD is getting SSE3 in the next revision or the one after that, continuing the tradition...
Click to expand...


you'd think that they would improve their current SSE2 engine first...
 
Originally posted by rayman2k2
you'd think that they would improve their current SSE2 engine first...
Click to expand...

Please forgive my ignorance, but I believe the Opteron is great at scalar SSE2, but packed SSE2, it seems less powerful than say, a Northwood-based P4.
 
Admittedly I'm not a programmer, but the only real shortcomings I've heard of AMD's SSE2 implementation is SSE2 likes clock speed.

There was something else tho... maybe SSE not SSE2...
 
the only real shortcomings I've heard of AMD's SSE2 implementation is SSE2 likes clock speed.
Click to expand...

I believe you and Josh are talking about the same thing - vectorized (packed) SSE2 performance is superior on the PIV as a function of clockspeed.

The Opteron/A64 architecture is apparently capable of more than one "FADD/FMUL per clockcycle" while the PIV is not, and so the Athlon 64 FX-51 is roughly twice as fast on Scalar SSE2 than the PIV 3.4EE, while being only 3/4ths as fast on vectorized SSE2.
 
Originally posted by leukotriene
I believe you and Josh are talking about the same thing - vectorized (packed) SSE2 performance is superior on the PIV as a function of clockspeed.

The Opteron/A64 architecture is apparently capable of more than one "FADD/FMUL per clockcycle" while the PIV is not, and so the Athlon 64 FX-51 is roughly twice as fast on Scalar SSE2 than the PIV 3.4EE, while being only 3/4ths as fast on vectorized SSE2.
Click to expand...

Interesting...

Good post :)
 
Originally posted by Merlin45
intel gets their sse2 performance from double clocking the math coproccesor.
Click to expand...

What? The 'math-coprocessor' is the FPU, the P4 double pumps the ALU...where on earth did you hear that?
 
Originally posted by Morley
What? The 'math-coprocessor' is the FPU, the P4 double pumps the ALU...where on earth did you hear that?
Click to expand...

Exacary... the P4's true x87 FPU is actually quite miserable when compared to the Athlon's ;)

Besides, the SSE registers are completely seperate from the FPU registers, unlike when using MMX.
 
Is that why similar performance (ie, 3200+ vs 3.2 P4) rated AMD CPU's will stomp a P4 during certain kinds of F@H operations?
 
Originally posted by xonik
I'm not sure what you are referring to.
Click to expand...


Originally posted by Josh_B
Please forgive my ignorance, but I believe the Opteron is great at scalar SSE2, but packed SSE2, it seems less powerful than say, a Northwood-based P4.
Click to expand...

sorry guys, should have been more clear. Opteron completely kills P4/Xeon in scalar SSE2, but it is vice versa for vectorized. I'm just saying, before they to jump to SSE3, they should try and improve the vectorized SSE2.

Then again, I'm not sure if SSE3 is a more advanced SSE2, meaning it is basically a SSE2 engine, but is more powerful.

I'm learning alot from this particular thread.
 
sry, I meant rapid execution engine, that is double clocked (and the alu's agu's and registers that comprise it) IIRC, sse2 instructions run through there, though I may be wrong. (I don't know what I was thinking when I wrote math coprocessor)
 
You must log in or register to reply here.
Back
Top