• A Great friend to the HardForum with a great kid that he is trying to get a scholorship to continue his schooling. Please give hime a vote! Only 24 hours left! Thanks.
    If you have an VOTE FOR KEENAN!

Intel Core i7-3820 Processor Review

H

HardOCP News

[H] News
Joined
Dec 31, 1969
Messages
0
Ryan and the gang at PC Perspective have thrown a little review lovin' on the Intel Core i7-3820 processor today. Quad core, 3.6GHz and 10MBs of L3 cache at $285? Could this be your next upgrade?

Those debating between the Core i7-3820 and the Core i7-2700K then have to decide if the added memory support and additional PCIe bandwidth are worth the added cost. If you are a heavy multi-GPU user then that might be the case, otherwise we are kind of at a standstill. Motherboards based on the Z68 platform available now should be able to support Ivy Bridge in the coming months but in truth we don't expect the processor portion of that upgrade to be particularly noteworthy - only the GPU side of the processor will see significant jumps. On the other hand, Ivy Bridge-E may or may not even actually use the LGA2011 socket, as Intel hasn't confirmed, though all indicators show that it will.
Click to expand...
 
Pretty interesting alternative to the i7-2600, if you don't mind paying more for a motherboard to get PCIe 3.0 and losing QuickSync.
 
Just makes me wonder about ivy bridge and would you rather wait for that or get this?
 
It's really too bad this isn't a K part.

If it were a K part it would be a great alternative to a 2700K/2600K.

I bought into X79 for the PCIe lanes. I didn't need the two extra cores and extra heat.

If this were a K part I would have been all over it.
 
Wait for ivy bridge. I won't be surprised if ivy does 5ghz easy and into possibly 6ghz with water. I'll wait.

The only thing that could prevent ivy from clocking high is if intel puts a much lower tjmax on it.
 
Galvin said:
Wait for ivy bridge. I won't be surprised if ivy does 5ghz easy and into possibly 6ghz with water. I'll wait.

The only thing that could prevent ivy from clocking high is if intel puts a much lower tjmax on it.
Click to expand...

Remember that die shrinks are not always that straight forward.

Intel is not only introducing a brand new 22nm process, but they are also introducing their new tri-gate transistors. Whenever new stuff like this is introduced there is a risk that something may not go as planned.

If die shrinks were that straight forward, the 32nm Llano CPU's would be the fastest, best overclocking CPU's AMD has right now, but in reality they're just about tying their 45nm predecessors.

Chances are good Ivy will be pretty awesome, I would just hold out until we hvae some tests in hand :p
 
Hurray for useless synthetic 3D benchmarks! Not that I'm expecting much of a variance, though...
 
Zarathustra[H];1038223417 said:
It's really too bad this isn't a K part.
Click to expand...
Did you actually read the reveiw?

"Unlike the original Sandy Bridge though, Sandy Bridge-E allows us to use "straps" or bus speed multipliers that still operate with the 3820."

By using a 125 MHz base clock you can make the multiplier limit a non-issue (that is you will most likely hit the chip's intrinsic limits first).
 
plugwash said:
Did you actually read the reveiw?

"Unlike the original Sandy Bridge though, Sandy Bridge-E allows us to use "straps" or bus speed multipliers that still operate with the 3820."

By using a 125 MHz base clock you can make the multiplier limit a non-issue (that is you will most likely hit the chip's intrinsic limits first).
Click to expand...

I've always been weary of bus overclocking. Too many subsystems affected that can randomly flake out in unpredictable ways.
 
Zarathustra[H];1038223739 said:
I've always been weary of bus overclocking. Too many subsystems affected that can randomly flake out in unpredictable ways.
Click to expand...

Spoiled with years of unlocked multipliers! Back in my day we used to tinker with ram voltage, timings, NB voltage in order to achieve overclocks, all the while hand-cranking a jerry-rigged power supply.
 
pelo said:
Spoiled with years of unlocked multipliers! Back in my day we used to tinker with ram voltage, timings, NB voltage in order to achieve overclocks, all the while hand-cranking a jerry-rigged power supply.
Click to expand...

Let's see.

In 1989 I overclocked my 286 using the bus speed. (There were no multipliers then)

In the early 90s I overclocked my 486 using bus speeds (there were multipliers then, but they were fixed. DX = 1x. DX2 = 2x and later DX4 = 4x)

In the mid to late 90s I overclocked my Pentiums using multipliers

In the early to mid 2000's I overclocked my Duron and various Athlons using multipliers (after unlocking them with pencils)

Then I stopped overclocking until 2010. Overclocked one Phenom II X4 using multipliers in 2010, and two X6's using multipliers in 2011. Also overclocked my new 3930K using turbo multipliers in 2011.

So I have done bus overclocks, but if I don't have to I usually just avoid them.
 
Last edited:
pelo said:
Spoiled with years of unlocked multipliers! Back in my day we used to tinker with ram voltage, timings, NB voltage in order to achieve overclocks, all the while hand-cranking a jerry-rigged power supply.
Click to expand...

Uphill in the snow right? :p
 
pelo said:
Spoiled with years of unlocked multipliers! Back in my day we used to tinker with ram voltage, timings, NB voltage in order to achieve overclocks, all the while hand-cranking a jerry-rigged power supply.
Click to expand...

Lets not forget dipswitches and jumpers in there too...
 
Zarathustra[H];1038223739 said:
I've always been weary of bus overclocking. Too many subsystems affected that can randomly flake out in unpredictable ways.
Click to expand...
Then you shouldn't have any complaint with the SNB-E overclocking because it doesn't work that way. The base clock strap only goes to the CPU clock, not any other components. With strap overclocking nothing else runs out of spec besides the CPU cores.
 
Zarathustra[H];1038223875 said:
Let's see.

In 1989 I overclocked my 286 using the bus speed. (There were no multipliers then)

In the early 90s I overclocked my 48c using bus speeds (there were multipliers then, but they were fixed. DX = 1x. DX2 = 2x and later DX4 = 4x)

In the mid to late 90s I overclocked my Pentiums using multipliers

In the early to mid 2000's I overclocked my Duron and various Athlons using multipliers (after unlocking them with pencils)

Then I stopped overclocking until 2010. Overclocked one Phenom II X4 using multipliers in 2010, and two X6's using multipliers in 2011. Also overclocked my new 3930K using turbo multipliers in 2011.

So I have done bus overclocks, but if I don't have to I usually just avoid them.
Click to expand...

Lol that was quite the "PWNED" right there :p
 
pxc said:
pfft. Don't forget soldering in new oscillators. ;) That was my first o/c in the 80s.
Click to expand...

No stranger to that one either...heck, I was soldering CPU sockets to use P3's in Celeron boards (BP-6).
 
Too wimpy of a cpu and I can blow more than $200 for a cpu I don't mind it I use it every day 365 days a year 24 hours a day.
 
pelo said:
Spoiled with years of unlocked multipliers! Back in my day we used to tinker with ram voltage, timings, NB voltage in order to achieve overclocks, all the while hand-cranking a jerry-rigged power supply.
Click to expand...

I'd be willing to bet anyone like Zarathustra[H] who has been on this forum 10+ years has prolly done a fair bit of non multiplier overclocking. But that's just a guess.... Feel free to join usin the decade + club.. in 9+ years. :p
 
pelo said:
Spoiled with years of unlocked multipliers! Back in my day we used to tinker with ram voltage, timings, NB voltage in order to achieve overclocks, all the while hand-cranking a jerry-rigged power supply.
Click to expand...

:D Oh man... Ain't that the truth. Not to mention breaking out the soldering iron and switching jumpers by hand. These damn snot nose kids these days! :p


^^^^
I need to read below before I reply. F'n troll N00b I am.
 
trick_m0nkey said:
Uphill in the snow right? :p
Click to expand...

Least somebody got the joke ;)

Yes, Zara, I too prefer unlocked multipliers. It's far easier and much more practical. But if you want to achieve the highest clock it still requires nearly all the same tinkering (minus the jumper that somebody mentioned. thank god that's mostly over). Though I doubt many of us go that far anymore. But since we're reminiscing...

http://www.motherboards.org/articles/guides/41_1.html
 
Then there I was with a magnifying glass a pin and rear window defroster repair kit connecting dots on my AMD to unlock it.
 
People are jumping to conclusions on the platform price compared to 2700K. 3820 is $50 cheaper as per 1k price ($285 as per anandtech and $332 as per intel ark). The motherboard argument is misleading: sure you can stick that 2700K in a $50 mobo, but will you? No you won't; compare like with like. ASUS Rampage IV Extreme vs Maximus IV Extreme-Z it's a $110 difference (that's with the Rampage's ridiculous tag). As it stands going 1155 you'll save a tiny bit on electricity, get the built-in GPU you're unlikely to care for and pocket maybe $50 on a $700 purchase. And the cheaper X79 boards will come, I've already googled pre-orders of Rampage IV Gene at $285 so I'm expecting these things to go for $260 when they actually will go on sale. Compare that to $190 Maximus Gene and you're right on the money. This only goes down as you move to value brands like Asrock and MSI.

If it wasn't for MicroCenter muddying the waters with unreal bundle deals I don't see why would anyone go for high end 1155 anymore. The main problem with MicroCenter for me is the fact they never carry the mobo I want. At this point it would be the Gene line and they only ever stock Extreme. I'm sure they're great and all but I'm really tired of the Washington Monument sized boxes at this point.
 
You must log in or register to reply here.
Back
Top