It kills me to say it, but I think my next build will be Intel...for the first time ever. Unless AMD pulls a miracle out next-gen, what Intel has done and keeps doing seals it - I can get roughly comparable performance in some tasks, but seeing this stomp on the FX-8150 at stock, and the 8150 at 4.8? No...I don't want to have to rewire my house to keep up with the power consumption AMD has now.
Plus AMD has essentially ceded the enthusiast market now...we'll see how PileDriver and SteamRoller go, but damn, AMD, you dropped the ball...No, you threw it down and smashed it on the pavement.
I wasn't even an AMD fan per se, they're just cheap - big gamble expecting Windows 8 to save your CPU when the level of adoption might not reach that of 7. Also annoyed that they couldn't make any hay out of Intel's mobo troubles, the company cares about graphics cards and that seems to be it.
This build is going into its 5th year and I never changed the CPU, couldn't be bothered, so Intel's socket change with every generation is only an issue if the board dies on you. If I want to save the extra cash it's good to know Ivy Bridge would give me another five years with maybe just a GPU change halfway through and whatever RAM I felt like adding at the time.
Those power consumption graphs really drive the nail home for Bulldozer - a 300W difference between IB overclocked at load and the 8150 just makes me wince! As fluxtatic says though, AMD seem to have given up competing at the top end.
In other news, I'd love to see bit-tech trying some old-school watercooling of just the cores, without the IHS!
"Overclocking
The concentration of heat with Ivy Bridge CPUs is proving to be a tricky obstacle to overcome. Air cooling yields much the same results as it did with Sandy Bridge 'K' Series CPUs with 4.6GHz usually an easy target. Using a multiplier of 50x, baseclock of 100MHz and a vcore of 1.4V, we pushed our Core i5-2500K to 5GHz with Extreme Load Line Calibration enabled and C-states disabled too. "
I'm also not sure about the power consumption numbers:
3570K at stock load only draws 5W less than the 3770K? HT has more of an effect than that....
also... the 3570K overclocked, draws more power than the 3770K OC? Again..... was expecting HT to have more of an effect.
And what about IB overcock on z68 boards? I've read about gigabyte problems (multiplier limit 39) but i have asus maximus 4 gene-z and was planning to ugrade for IB and descent LC setup... i've run quick trough the article and can't get enough info on this
send from my BLADE via tapatalk
For the locked multiplier issue, to be honest, it is the Award BIOS limitation for our Z68 boards to Overclock with IVB K series CPU. It will be locked at 39." - Gigabyte USA
No one else except Gigabyte uses Award BIOS for their Sandy Bridge boards.
Have you taken the lid of the CPU, as Overclock found that their CPU sample used TIM rather than solder, potentially explaining the difference in temperatures, combined with the higher power density.
Originally Posted by K404 Typo: 2500K should read 3570K
"Overclocking
The concentration of heat with Ivy Bridge CPUs is proving to be a tricky obstacle to overcome. Air cooling yields much the same results as it did with Sandy Bridge 'K' Series CPUs with 4.6GHz usually an easy target. Using a multiplier of 50x, baseclock of 100MHz and a vcore of 1.4V, we pushed our Core i5-2500K to 5GHz with Extreme Load Line Calibration enabled and C-states disabled too. "
I'm also not sure about the power consumption numbers:
3570K at stock load only draws 5W less than the 3770K? HT has more of an effect than that....
also... the 3570K overclocked, draws more power than the 3770K OC? Again..... was expecting HT to have more of an effect.
Don't forget the 3570K was overclocked to a higher frequency and with much more voltage. We haven't noticed massive increases with HT/no HT - the 2500K and 2600K have less than 20W difference, albeit that's much more than the 3570K and 3770K.
and me sat here with a 720BE never having found a reason to upgrade! Look, if I'm thinking about buying a hybrid car, why would I ever continue to buy AMD? Anything over 60fps on 1080p is a waste because I can't see it, I want a PC that's quiet, competent and cheap. And that is probably going to be Intel powered...
As an i7 930 user I wasn't getting too excited about this, until I got to the power consumption - the figures are so much lower than what I'm getting right now, I think it'll almost be worth upgrading for the electricity bill...Almost. Still waiting for a massive performance jump from the 930 but I think I could be waiting a long time... Most of the gaming I do is GPU limited so an upgrade there makes more sense for me...maybe Haswell will be the one that makes me take the jump; I don't think it'll be anything AMD though, barring a minor miricle.
Originally Posted by warejon9 Have you taken the lid of the CPU, as Overclock found that their CPU sample used TIM rather than solder, potentially explaining the difference in temperatures, combined with the higher power density.
That was shot down when some enthusiast out there delidded his 3770k and ran it with his cooler making direct contact on the die. Almost no difference in temps. Intel has also come out and said that the reason is thermal density. You can find the thread about this somewhere in the Hardware section of the forums.
My 3570k should be waiting for me when I get home. It's gonna be an enormous upgrade from my i5-750.
Originally Posted by faugusztin @Bede: there is a relatively big issue with that - without the IHS the chip is not fixed inside the LGA socket.
This is a modding site, I'm sure someone could find a way. It's not the IHS that fixes it in, it's the pressure from the contact plate - make a new contact plate like the pre-IHS ones and the problem could possibly be solved. Not an easy task, nor something that many would be comfortable experimenting with I agree.
Originally Posted by Neogumbercules That was shot down when some enthusiast out there delidded his 3770k and ran it with his cooler making direct contact on the die. Almost no difference in temps. Intel has also come out and said that the reason is thermal density. You can find the thread about this somewhere in the Hardware section of the forums.
My 3570k should be waiting for me when I get home. It's gonna be an enormous upgrade from my i5-750.
The bottleneck could still be die area (unrelated to process density): without testing a soldered IHS ivy bridge to a TIM-only ivy bridge, it's not possible to tell if the major barrier to heat conduction is the thermal resistance of the boundary between the silicon and the heat spreader/sink, or the area of the thermal boundary (I.e. w/k vs w/mm^2). The soldered IHS has such a low thermal resistance that the conduction area is many times that of the bare die.
Well it depends. Are you over clocking? If you are and want more sane temperatures, for now IB doesn't sound like the way to go.
If you aren't overclocking or were looking at a mild overclock or the temperatures being sky high don't bother you, than IB might be the bet. If you are upgrading from SB, then the answer really depends on
1) Do you need to possibly eek just a couple percentage points more performance OC'd out of the processor, or if stock looking for maybe a 10% gain.
2) Do you rely on the IGP? If so, then deffinitely upgrade.
3) Do you want/need integrated USB3? Do you want/need PCIe 3.0?
If anything above doesn't apply, than going from SB to IB makes absolutely no sense. Frankly I wouldn't consider it if I had a SB. But I don't, I have an E7500, so I'll be getting a 3570 or 3570k sometime in June once hopefully board and processor prices have come down slightly, any potential major glitches are found and I can save just a wee bit more money.
I am a once every 3 or so year upgrade kind of guy (and when I upgraded 3 years ago it was from a P4 2.53 to my E7500, not exactly a new processor then). I am sure the next time I upgrade will be to something on the 14nm process, probably the Tock of the 14nm.
Almost pulled the trigger on one today but found all the motherboard variants a bit confusing and gave up :o Just want to overclock and run SLI without paying 200quid for a mobo, more like 100.
I'm lucky enough to have a 2600K capable of performing as below so Ivybridge CPU upgrade for me???
Nah
Looking at the test figures my current CPU is capable of outperforming both the tested 3570K and 3770K; further still it doesn't require a yet to be designed cooling solution to prevent it going thermonuclear at high overclocks...
Comments 1 to 25 of 43
ReplyPlus AMD has essentially ceded the enthusiast market now...we'll see how PileDriver and SteamRoller go, but damn, AMD, you dropped the ball...No, you threw it down and smashed it on the pavement.
This build is going into its 5th year and I never changed the CPU, couldn't be bothered, so Intel's socket change with every generation is only an issue if the board dies on you. If I want to save the extra cash it's good to know Ivy Bridge would give me another five years with maybe just a GPU change halfway through and whatever RAM I felt like adding at the time.
In other news, I'd love to see bit-tech trying some old-school watercooling of just the cores, without the IHS!
"Overclocking
The concentration of heat with Ivy Bridge CPUs is proving to be a tricky obstacle to overcome. Air cooling yields much the same results as it did with Sandy Bridge 'K' Series CPUs with 4.6GHz usually an easy target. Using a multiplier of 50x, baseclock of 100MHz and a vcore of 1.4V, we pushed our Core i5-2500K to 5GHz with Extreme Load Line Calibration enabled and C-states disabled too. "
I'm also not sure about the power consumption numbers:
3570K at stock load only draws 5W less than the 3770K? HT has more of an effect than that....
also... the 3570K overclocked, draws more power than the 3770K OC? Again..... was expecting HT to have more of an effect.
send from my BLADE via tapatalk
No one else except Gigabyte uses Award BIOS for their Sandy Bridge boards.
Don't forget the 3570K was overclocked to a higher frequency and with much more voltage. We haven't noticed massive increases with HT/no HT - the 2500K and 2600K have less than 20W difference, albeit that's much more than the 3570K and 3770K.
That was shot down when some enthusiast out there delidded his 3770k and ran it with his cooler making direct contact on the die. Almost no difference in temps. Intel has also come out and said that the reason is thermal density. You can find the thread about this somewhere in the Hardware section of the forums.
My 3570k should be waiting for me when I get home. It's gonna be an enormous upgrade from my i5-750.
This is a modding site, I'm sure someone could find a way. It's not the IHS that fixes it in, it's the pressure from the contact plate - make a new contact plate like the pre-IHS ones and the problem could possibly be solved. Not an easy task, nor something that many would be comfortable experimenting with I agree.
Indeed it should be - the move from an i5-760 to an i5-2500K was big enough. Enjoy!
Those temps look scarey, I wonder if intel will change the heatspreader assembly process to ensure a more effective platform for coolers?
Also, did I miss this, or are reviewers genuinely avoiding posting the integrated GPU performance?
If you aren't overclocking or were looking at a mild overclock or the temperatures being sky high don't bother you, than IB might be the bet. If you are upgrading from SB, then the answer really depends on
1) Do you need to possibly eek just a couple percentage points more performance OC'd out of the processor, or if stock looking for maybe a 10% gain.
2) Do you rely on the IGP? If so, then deffinitely upgrade.
3) Do you want/need integrated USB3? Do you want/need PCIe 3.0?
If anything above doesn't apply, than going from SB to IB makes absolutely no sense. Frankly I wouldn't consider it if I had a SB. But I don't, I have an E7500, so I'll be getting a 3570 or 3570k sometime in June once hopefully board and processor prices have come down slightly, any potential major glitches are found and I can save just a wee bit more money.
I am a once every 3 or so year upgrade kind of guy (and when I upgraded 3 years ago it was from a P4 2.53 to my E7500, not exactly a new processor then). I am sure the next time I upgrade will be to something on the 14nm process, probably the Tock of the 14nm.
good to know. Just what I was interested in learning. +rep ;)
Now.. Just to replace my 4870!
Nah
Looking at the test figures my current CPU is capable of outperforming both the tested 3570K and 3770K; further still it doesn't require a yet to be designed cooling solution to prevent it going thermonuclear at high overclocks...
World's most posted picture... :D:D:D:D:D
And rightly so :D :D :D :D :D ;) :)
-
« Previous
-
1
-
2
-
Next »
Discuss in the forums