Due to Series S/X alignment requirements, Baldur’s Gate 3 will release on PS5 first

[Remij 4,687]

1 hour ago, jehurey said:

No that's not a "higher number"

 

Because compute units is not a performance metric.

 

That's like you saying "my car is faster than yours"

 

And I respond "oh really, how so"

 

And you say "well, my engine has 8 cylinders"

 

That doesn't mean shit. That isn't a performance metric.

 

You, after already being told that teraflops isn't really indicative of real-world performance...........tried to throw out a statistic that MEANS NOTHING EXCEPT FOR THE FACT THAT ITS USED TO DETERMINE WHAT, EXACTLY???????????? Teraflops.

 

And you thought you could slip that one by.

 

He's trying to argue whether 12 is a higher number than 10  

LOL he just said compute units aren't a performance metric... when they are literally one of the components (the primary component) in measuring the theoretical max performance of a chip.. that's literally why the metric exists moron..  

 

You ABSOLUTELY need to calculate the number of compute units to figure out the maximum theoretical performance of a chip... and the higher the number the more compute performance. PERIOD.  Within architectures, it absolutely tells you the compute capabilities of one chip compared to another.  Xbox has more compute capability than PS5. PERIOD.

 

And I ALREADY told your bitch ass that there is more to real world performance than just compute.... you're not telling anyone anything sweetie  

 

 

 

[jehurey 3,241]

1 hour ago, Remij said:

 

He's trying to argue whether 12 is a higher number than 10  

LOL he just said compute units aren't a performance metric...

 

 

 

No they're not.

 

They used to CALCULATE a performance metric.

 

..........and calculation is made and given in WHAT terminology, again??????

 

(he's not going to answer that question, because it proves me correct about what he's doing)

 

How fast is your car?

 

Its eight cylinders fast........eight cylinders is faster than six cylinders.

[Remij 4,687]

On 2023-08-06 at 3:25 PM, jehurey said:

No they're not.

 

They used to CALCULATE a performance metric.

 

..........and calculation is made and given in WHAT terminology, again??????

 

(he's not going to answer that question, because it proves me correct about what he's doing)

 

How fast is your car?

 

Its eight cylinders fast........eight cylinders is faster than six cylinders.

LOL look at this fucking idiot... he quotes me... cut's off the quote... and then proceeds to "correct me" by saying exactly what I said in the part he cut off  

 

Quote

LOL he just said compute units aren't a performance metric... when they are literally one of the components (the primary component) in measuring the theoretical max performance of a chip.. that's literally why the metric exists moron..  

 

What a dumbass    

 

[jehurey 3,241]

43 minutes ago, Remij said:

LOL look at this fucking idiot... he quotes me... cut's off the quote... and then proceeds to "correct me" by saying exactly what I said in the part he cut off  

 

 

What a dumbass    

 

Nope..........you don't finish your OWN sentence

 

Let's take a look:

 

Quote

LOL he just said compute units aren't a performance metric... when they are literally one of the components (the primary component) in measuring the theoretical max performance of a chip..

is theoretical max performance..............measure in compute units?

 

Or something else?

 

And that would be????????

 

he already knows I'm corning him...........he already knows he can't get out of it................he's just trying to pad out his non-answer

[Remij 4,687]

1 hour ago, jehurey said:

Nope..........you don't finish your OWN sentence

 

Let's take a look:

 

is theoretical max performance..............measure in compute units?

 

Or something else?

 

And that would be????????

 

he already knows I'm corning him...........he already knows he can't get out of it................he's just trying to pad out his non-answer

You're too stupid...   I never said it was measured in compute units... I said compute units factor into the measurement... OMFG  

 

Theorectical max performance is CALCULATED by taking the number of shader cores per CU and multiplying that by how many CUs there are, and then multiplying that by the max clock speed... then multiplying that by how many FP32 instructions per clock it can perform.

 

For Nvidia substitute CUs for cuda cores per SM x the number of SMs x clock speed x instructions per clock FP32..

 

Clown  

Edited August 8, 2023 by Remij

[jehurey 3,241]

1 hour ago, Remij said:

You're too stupid...   I never said it was measured in compute units...

but you are using it to boast about it in terms of performance.

 

That it CONTRIBUTES to something performance-related. That you, coincidentally, don't want to clarify how that is quantified.

 

Dude.............I can easily shut down EVERY SINGLE angle you try to take............every time.

 

There's no way you're getting out of this, I can easily corner you.

 

You just listing something about the hardware specs...................JUST BECAUSE?

[Remij 4,687]

15 minutes ago, jehurey said:

but you are using it to boast about it in terms of performance.

 

That it CONTRIBUTES to something performance-related. That you, coincidentally, don't want to clarify how that is quantified.

 

Dude.............I can easily shut down EVERY SINGLE angle you try to take............every time.

 

There's no way you're getting out of this, I can easily corner you.

 

You just listing something about the hardware specs...................JUST BECAUSE?

 

You're such a tool dude  

 

The number of CUs or SMs or "shader cores".... DOES contribute to something performance related moron.... the compute performance of the chip  

 

And what did I say about compute performance?  .... That it's not the only aspect of overall performance.  AMD's RDNA2 and RDNA3 products demonstrate this perfectly.  In RDNA3 each shader core inside each CU can perform dual instructions which THEORETICALLY allows the chip to perform 2x faster in those operations... and yet RDNA3 GPUs aren't 2x faster in games than RDNA2... are they?    RDNA3 TFs don't equal that of RDNA2... and that's because despite giving each shader core the ability to perform dual instructions, there's not 2x the amount of CUs.. so those shader cores don't have access to 2x the L1 cache, 2x the ROPs.. ect ect..  So while they may perform SOME operations faster, not everything will be.  Then you have other aspects like memory bandwidth and so on...

 

So while it isn't an excellent indicator for OVERALL performance... because there are other aspects of rendering which don't rely on, or benefit from compute being able to perform 2x the instructions per clock.. it DOES indicate the COMPUTE performance of the chip... and can be used to compare products within that same RDNA3 generation with each other.

 

Jesus christ.. this is so fucking embarrassing for you  

[Ike★ 2,919]

Series S was a mistake.  

[Goukosan 2,258]

4 hours ago, Ike said:

Series S was a mistake.  

 

Anyone with a brain could see that coming... Except lemmings of course

[jehurey 3,241]

16 hours ago, Remij said:

 

You're such a tool dude  

 

The number of CUs or SMs or "shader cores".... DOES contribute to something performance related moron.... the compute performance

as measured in?

 

I can checkmate him..........every..........time.

 

he knows it which is why he's trying to stall for time by copy/pasting wikipedia passages

[Remij 4,687]

4 hours ago, jehurey said:

as measured in?

 

I can checkmate him..........every..........time.

 

he knows it which is why he's trying to stall for time by copy/pasting wikipedia passages

 

On 2023-08-06 at 12:12 AM, Remij said:

I'm not going to feel anything about that... Teraflops haven't been an accurate representation in the PC space since forever ago... like as if Nvidia and AMD TFs were ever comparable..  

 

Xbox has more compute capability... that much is 100% fact.  Not all games will be compute bound, however... 

 

Moron.

[jehurey 3,241]

15 hours ago, Remij said:

 

 

Moron.

I got him to defend teraflops

 

That's basically him caving in and giving up.

 

Thanks for doing and now admitting to doing EXACTLY what I said you were way back about 15 posts ago.

 

Keep on talking about "capabilty"

 

The PS5 is going ahead and demonstrating actual "ability". And Digital Foundry is saying that the capability argument isn't worth shit.

 

What are you going to do????? Respond back with typing "dAh ToOlZ!!!!" Post?

 

Hey Gouko, Iwonder if I can make Lemij start whining about "da toolz" again. I probably can.