Some New AMD Carrizo Details

[Opcode]

Am I denying it? I'm saying putting more lets say disabled cores to get a bigger die size is not going to make a noticeable difference to your temperatures, as I proven above a bigger die size doesn't do anything to your temperatures.

My dispute isn't about die size at all (I haven't mentioned it once). The fact of the matter is if you got cores sitting there doing nothing they can mitigate some of the heat from the sounding cores. As they are a semiconductor. Curious which cores did you disable on your chip when you ran it as a quad core?

 

that would definitely make some money for them 

It may be quite an appealing product to have one of the strongest consumer SoC's on the planet. Tho the price for them would be quite a hurdle. Especially if something goes wrong and a CPU core dies or another then you have to replace the entire chip. Technology is headed in that direction tho for right now a gamble like that could prove fatal for the company. I personally would like to have one but I guarantee there's a lot of people out there who don't.

[Faa]

My dispute isn't about die size at all (I haven't mentioned it once). The fact of the matter is if you got cores sitting there doing nothing they can mitigate some of the heat from the sounding cores. As they are a semiconductor. Curious which cores did you disable on your chip when you ran it as a quad core?

You can't chose which cores you want to disable, I think it's just core#0, core#1, core#2, core#3. X99 has the choice to disable/enable any core.

[Opcode]

You can't chose which cores you want to disable, I think it's just core#0, core#1, core#2, core#3. X99 has the choice to disable/enable any core.

They are labeled to correspond with a particular core it's just finding out which. According to Intel they run like this for your chip.

 

Core 0 --- Core 1

Core 2 --- Core 3

Core 4 --- Core 5

[vm'N]

With whatever amount of IPC increase over Bdver3 validations on SiSoft are putting Carrizo around the GFLOPS of a mobile Sandy Bridge i3 @ 2.0 GHz while at 15w TDP. While not competing clock for clock (not important) the power consumption could show impressive in comparison. With Intel not yielding any major IPC increases from Sandy Bridge all the way through Broadwell. Carrizo is shaping up to possibly be a bigger success in the mobile market than any of their past mobile generations. Intel undoubtedly still easily has the upper hand tho it's nice to finally possibly nice see an AMD CPU capable of competing within range of Sandy Bridge in performance at much lower power consumption. The 3.4 GHz I would imagine is only effective within thermal limits (iGPU would have to be idle).

GFLOP != actual performance.

AMD was to late for the mobile market. This seems like a attempt to regain mobile market share.

Their previous mobile line-up have been a huge fiasco, to be honest.

AMD will most likely continue with competing on price.

My guess is that the sustainable turbo will be around the ~3GHz.

[LukaP]

They are labeled to correspond with a particular core it's just finding out which. According to Intel they run like this for your chip.

 

Core 0 --- Core 1

Core 2 --- Core 3

Core 4 --- Core 5

How do you have the patience... I applaud you. Just know that he is a known troll who has been proven wrong countless times (plus a few times in this thread) and that you will never be able to convince him youre right, even though you clearly are....

[FatalityDiablo]

lets hope it does not perform like it tastes..

[Opcode]

GFLOP != actual performance.

AMD was to late for the mobile market. This seems like a attempt to regain mobile market share.

Their previous mobile line-up have been a huge fiasco, to be honest.

AMD will most likely continue with competing on price.

My guess is that the sustainable turbo will be around the ~3GHz.

Of course, tho it's nice to see AMD competing within margin of Sandy Bridge at least in FFT's and doubling its performance in matrix multiplication. Meaning double precision performance has at least landed on Sandy Bridge levels. I applaud them (so far) with what they have done with Carrizo. With floating point performance up all they need to do is deliver on integer performance. If priced accordingly (unlike the FX-7600P and FX-7500) they may actually sell a few of these units. I personally would like to see them push out a few 4k laptops with Carrizo at its heart for around the $600 price point.

 

The ES chip apparently has a base clock of 1.8 GHz with a 3.4 GHz turbo state (within thermal margin).

[vm'N]

Of course, tho it's nice to see AMD competing within margin of Sandy Bridge at least in FFT's and doubling its performance in matrix multiplication. *Meaning double precision performance has at last landed on Sandy Bridge levels. I applaud them (so far) with what they have done with Carrizo. **If priced accordingly (unlike the FX-7600P and FX-7500) they may actually sell a few of these units. I personally would like to see them push out a few 4k laptops with Carrizo at its heart for around the $600 price point.

 

The ES chip apparently has a base clock of 1.8 GHz with a 3.4 GHz turbo state (within thermal margin).

* The problem is that doesn't tell much about actual performance.

 

** Carrizo, will increase AMDs sales in the laptop segment, the problem is that carrizo, will most likely continue competing on price.

OEMs will most likely build it around other cheap components. I too would find a cheap 4k laptop with carrizo APU interesting.

[Faa]

How do you have the patience... I applaud you. Just know that he is a known troll who has been proven wrong countless times (plus a few times in this thread) and that you will never be able to convince him youre right, even though you clearly are....

http://www.pugetsystems.com/labs/articles/Core-i7-5960X-vs-4960X-Performance-Comparison-588/page3

Measuring temps from the hottest core and as we know a single 4960x core consumes more power than the rest, using the board as fan control dropping variance in there, 5930K having a bigger die size since it's physically an octocore, temps rather being within margin of error, proves that a bigger die size does nothing.

I've asked him to prove a bigger die size will affect your temperatures, all I've gotten is they can act as a passive heatsink which is not proving his claim. The cooler I used was a H100, just google the backplate issues like most coolers these days, something you don't use on lga2011, the contact is just much better and that's the reason why I saw a huge difference. Besides he's the guy who made claims that Nvidia stole Mantle code and putted in their own DX11 drivers and arguing with his knowledge against an article from MS stating DirectX has been low-level since the 8th version.

[Opcode]

* The problem is that doesn't tell much about actual performance.

 

** Carrizo, will increase AMDs sales in the laptop segment, the problem is that carrizo, will most likely continue competing on price.

OEMs will most likely build it around other cheap components. I too would find a cheap 4k laptop with carrizo APU interesting.

Indeed, I myself am more interested in integer performance as that's all that matters in software and gaming.

 

http://www.pugetsystems.com/labs/articles/Core-i7-5960X-vs-4960X-Performance-Comparison-588/page3

Measuring temps from the hottest core and as we know a single 4960x core consumes more power than the rest, using the board as fan control dropping variance in there, 5930K having a bigger die size since it's physically an octocore, temps rather being within margin of error, proves that a bigger die size does nothing.

I've asked him to prove a bigger die size will affect your temperatures, all I've gotten is they can act as a passive heatsink which is not proving his claim. The cooler I used was a H100, just google the backplate issues like most coolers these days, something you don't use on lga2011, the contact is just much better and that's the reason why I saw a huge difference. Besides he's the guy who made claims that Nvidia stole Mantle code and putted in their own DX11 drivers and arguing with his knowledge against an article from MS stating DirectX has been low-level since the 8th version.

You're still not understanding it at all. Hottest cores are irrelevant to coolest cores. You need to dig up proof that the two cores parked next to the disabled cores are running hotter than the rest of them. That's the only way you'll be able to prove that cores do not act as passive sinks. Also are you still going on about that rumor? I only posted it and I did not endorse it (as said in the post). You can try your best to discredit me with it as it wont happen (as proven time and time again).

[Faa]

You're still not understanding it at all. Hottest cores are irrelevant to coolest cores. You need to dig up proof that the two cores parked next to the disabled cores are running hotter than the rest of them. That's the only way you'll be able to prove that cores do not act as passive sinks. Also are you still going on about that rumor? I only posted it and I did not endorse it (as said in the post). You can try your best to discredit me with it as it wont happen (as proven time and time again).

I don't need to prove those cores don't act as a passive heatsink since I'm not denying it. Out of the 6 cores there's always a core that runs the hottest, which one it is and by how much varies from chip to chip.

Besides seems like you are struggling to understand it by yourself, the 6 cores on the 5930K are passively cooled off by the ones that are disabled. 5930K is physically an octocore, the 4960x is smaller because it's physically a hexacore. My 3930K is physically an octocore, we both would agree core#5 is closer to the disabled pair right?

Core#5 (which is closer to the disabled ones) is running 4° warmer;

You backed your rumour up with "my own sources" so it was pretty much you who was starting that rumour and I can't see anyone else making that claim yet and you made the claim after you I provided you the proof that Nvidia's DX11 is nearly on par with Mantle in terms of cutting CPU overhead.

Also, weren't you the guy who was claiming Carrizo will kick the living shit out of Sandy Bridge? Now you're just "Gflops is roughly the same". Carrizo will come out and your knowledge failed you again. Recently you made the claim when FX CPU's were released they were ahead of the competition at that time. Afaik you made the claim you don't need scalars for variable refresh rate on desktop monitors. You're not difficult to disprove if your knowledge = AMD PR.

[Opcode]

I don't need to prove those cores don't act as a passive heatsink since I'm not denying it. Out of the 6 cores there's always a core that runs the hottest, which one it is and by how much varies from chip to chip.

Besides seems like you are struggling to understand it by yourself, the 6 cores on the 5930K are passively cooled off by the ones that are disabled. 5930K is physically an octocore, the 4960x is smaller because it's physically a hexacore. My 3930K is physically an octocore, we both would agree core#5 is closer to the disabled pair right?

Core#5 (which is closer to the disabled ones) is running 4° warmer;

You backed your rumour up with "my own sources" so it was pretty much you who was starting that rumour and I can't see anyone else making that claim yet and you made the claim after you I provided you the proof that Nvidia's DX11 is nearly on par with Mantle in terms of cutting CPU overhead.

Also, weren't you the guy who was claiming Carrizo will kick the living shit out of Sandy Bridge? Now you're just "Gflops is roughly the same". Carrizo will come out and your knowledge failed you again. Recently you made the claim when FX CPU's were released they were ahead of the competition at that time. Afaik you made the claim you don't need scalars for variable refresh rate on desktop monitors. You're not difficult to disprove if your knowledge = AMD PR.

You would also have to prove which cores were disabled as it can be any two that are the weakest (it could be binned with any two bad cores). My argument is based on cores acting as passive sinks which you're now agreeing with me on so my work is done here. Is the disabled or low powered cores being parked next to active cores going to make a drastic difference in load temperatures? No. Will they help with thermal dissipation? Yes. As stated AMD uses the concept in their APU's for holding longer boost clocks before pushing the thermal threshold. This will help the iGPU hold higher frequencies (PowerTune) with heat is being migrated elsewhere (cores -> IHS surface area).

 

I never back up any rumors as that's one of the most stupid things to do to get yourself caught up in speculation. The very first post that you seemed to have a problem with I even said in my own words that I don't endorse the claim by any means (I don't buy it). Tho you still like to think I am wrong some how because I copy and pasted the rumor to stir up speculation. I know for a fact that it doesn't take much to step on your toes and trigger your inner fanboy. In that extent if you want to continue to believe that Nvidia stole Mantle code then cool. That's your own opinion (even tho it's wrong) but it doesn't retain to any aspect of this thread. I and everyone else would label that as thread derailing. If you have a personal vendetta with me (I know you do) shoot me a message instead of playing the e-peen game of who's right and wrong on the forum publicly as it doesn't contribute to anything other than self ego.

 

I never made that claim. If you can find a post of my exact words "kicking the living shit out of Sandy Bridge" feel free to quote me. It seems like every time I prove you wrong that you like to derail the the thread and the very subject of your own posts to try and bend reality into your favor. It doesn't work like that and you need to accept that you're more wrong than right most of the time (especially between us). I don't mind continuing the next several years doing what I do when it comes to you because it only builds my reputation for standing against someone so hard headed. You either know what you know or you don't know nothing. This is why I hate using outside sources for my own knowledge. Tho after people like you complain about it I provide them and they are concrete. Meanwhile you still seem to got a problem over me being right off the top of my head while you dig through Google for unreliable sources claiming the contrary.

 

This thread is about Carrizo. If you don't want to talk about Carrizo then leave because this is not the place to talk about the core temperatures of your 3930k.

[patrickjp93]

How do you have the patience... I applaud you. Just know that he is a known troll who has been proven wrong countless times (plus a few times in this thread) and that you will never be able to convince him youre right, even though you clearly are....

Wait wait wait, hold one... @Opcode is right about something?! I, with the power vested in me by our divine overlords, will be the judge of that. Mere mortals should not presume such a thing. This is a cosmic matter.

[Opcode]

Wait wait wait, hold one... @Opcode is right about something?! I, with the power vested in me by our divine overlords, will be the judge of that. Mere mortals should not presume such a thing. This is a cosmic matter.

I'm not religious.  :rolleyes:

 

Let teh Aliens roll in. 

[Faa]

You would also have to prove which cores were disabled as it can be any two that are the weakest (it could be binned with any two bad cores). 

I didn't turn any cores off >.< Like I told you the 3930K is physically an octocore with 2 of them being disabled so 6 can be only up. I just proven the core that's the closest to the disabled one was running warmer than the core which was the furthest from it.

 

 

 

I never made that claim. If you can find a post of my exact words "kicking the living shit out of Sandy Bridge" feel free to quote me. It seems like every time I prove you wrong that you like to derail the the thread and the very subject of your own posts to try and bend reality into your favor. 

http://linustechtips.com/main/topic/279104-prove-amds-superiority-to-me/page-5#entry3800738

Not exactly, the expression "with flying colors" is the same thing.

[Opcode]

I didn't turn any cores off >.< Like I told you the 3930K is physically an octocore with 2 of them being disabled so 6 can be only up. I just proven the core that's the closest to the disabled one was running warmer than the core which was the furthest from it.

I don't see any proof of this warmer core running closest to a disabled core. Like said the core configuration on your chip could be scrambled with specific bad cores disabled. Without actual proof that X core is running next to Z core you're only making random guesses.

 

http://linustechtips.com/main/topic/279104-prove-amds-superiority-to-me/page-5#entry3800738

Not exactly, the expression "with flying colors" is the same thing.

30% IPC increase over Bdver3 would put Bdver4 ahead of Sandy Bridge. That would be "with flying colors" in my book. So far floating point performance is knocking on Sandy Bridges back door tho the real numbers will come from when we find out integer performance.