Sandy Bridge Reviews Released!

I thought these might be dropping early after people picked up chips over the weekend
 
Observation for users doing Virtualization.

According to Anandtech,

The overclocking friendly K-series Sandy Bridge 2500K and 2600K VT-x. Do not support VT-d
VT-d however is present on non-K series 2400/2500/2600 VT-x/VT-d
i5 2300 has VT-x but no VT-d
i3-2100 and i3-2120 no VT-x / no VT-d

Edit : I saw 2 different listings, Anandtech as listed above, but Semiaccurate shows i3-2100 and i3-2120 both having VT-x. Apology if wrong info.
Edit : still no VT-x according to Anandtech for 2100 / 2120
 
Last edited:
Looks like the IPC gains are enough to put the 2600k trading blows with the 980x at stock clocks, impressive stuff.
 
People were reporting picking up 2600k's at a couple Frys and Canada Computer (I think thats the name) locations. Lucky bastards.

my Fry's is about 30 mins away, I hate driving in the East Bay though. Guess I'll wait till they show up on NewEgg, unless they pull their standard price gouging BS
 
Tech reaction hit 5.1ghz with 1.45v while overclock 3d hit 5ghz at 1.55v.

What is max spec voltage on Intel's 32nm?
 
Observation for users doing Virtualization.

According to Anandtech,

The overclocking friendly K-series Sandy Bridge 2500K and 2600K VT-x. Do not support VT-d
VT-d however is present on non-K series 2400/2500/2600 VT-x/VT-d
i5 2300 has VT-x but no VT-d
i3-2100 and i3-2120 do not support VT-x/VT-d

So what will the absense of VT-d cause for VMs in VMWare workstation? I believe VT-x is what is required for installation of 64 bit virtual OS's.

I'm trying to decide if this is going to affect my decision.
 
You guys hear something? I do............


f316F.jpg
 
Very sweet indeed. Bonus that apparently we can use good old CAS 9 1333mhz memory and performance isn't hindered... reviews are suggesting no new major memory requirements from any current Core i builds.
well with the vast majority of the overclocking being directly derived from multiplier increases, there won't be any issue with slower memory having to be overclocked and/or run at a divider (like when you increase bclk) While that means performance isn't "hindered" by older slower memory, it does take one of the avenues to increase system performance out of the equation.
 
I know the launch is this week but any word on what day? I have seen both the 5th and the 9th. The 9th seems doubtful since its a Saturday. I know Intel is scheduled the 5th at Ces.
 
Regarding the anandtech review
1 The chart shows the GPUs to have different xter counts. WTF?
2: 4.1Ghz turbo is talked about in the very first paragraph, but the charts and images show 3.8Ghz max.
3: While the vCore numbers concern me, I'm still drooling for a [email protected]+Ghz.
:)

You guys hear something? I do............
You must be imagining things.
 
Last edited:
So what will the absense of VT-d cause for VMs in VMWare workstation? I believe VT-x is what is required for installation of 64 bit virtual OS's.

I'm trying to decide if this is going to affect my decision.

Hi, currently (I suppose this one or two years, not sure)

VT-d is consideration for people heavy into virtualization. By expectation we should buy Xeon server build, but many could only afford core i desktop setup.

For example, the current hot topic in data storage forum about building NAS, one possible scenario is such that the NAS system disk runs on Virtual disk, but use VT-d to pass a physical Storage Controller (with many physical disks linked directly to it) to the NAS VM. In this way, you gain the flexibility of NAS OS in VM, (thus all benefits of doing so), yet expose the physical disks to it (thus all the benefits of raw disks with minimal interference from the virtualization platform layer)

Xen has some interesting examples with user doing unique setups, (like passing multimedia PCI-card to VM). I read that a while back but have no real experience with it cause I am not doing Xen and current test machine does not support VT-d

I am not sure about VMware Workstation but ESXi I believe can do pass through. Myself using VMware Player on home machine with no VT-d support.

BTW, database, same theory, you can setup DB server OS/SW on virtual disk, and use VT-D to pass additional storage RAID controller (not used by host platform) with example, say 8 physical disks as database storage to the SQL Server. In this way, the critical data area can be on something more familiar, instead of bunch of VMDK files. However, it is true if this scenario is real SQLServer one should go for Xeon/ECC RAM instead. :)

Cheers
 
Last edited:
The 2600k is about equal to the 980x clock for clock, does 4.8-5ghz, and according to the bit-tech article it draws the same power at 4.85ghz that a stock i7 950 does at 3.06ghz.

I would say Intel delivered.
 
Canada Computers in Toronto has been selling the procs for almost a week for $317 CAN.
 
The 2600k is about equal to the 980x clock for clock, does 4.8-5ghz, and according to the bit-tech article it draws the same power at 4.85ghz that a stock i7 950 does at 3.06ghz.

I would say Intel delivered.

I bet ppl who bought a 980x are kicking themselves what a waste 900bucks:cool:
 
So I have a Q9550 @ 3.4GHz - am I still good for a while?
 
^ do the games you play run alright (in your opinion)? i'm upgrading because my Q6600 is not up to the task for some of my favorite games.
 
I'm curious what the voltage spec limit for these is, I thought 32nm was somewhere near 1.38v. Most of the reviewers used 1.45v to oc, only Bit-tech did a solid 4.85ghz at 1.35.
 
Just a quick side tangeant... which mobo manufacturers would you recommend for a paired mobo with these CPUs? I've had problems with some lately... so I'm cautious. I think it's important to pair the CPU with a good MOBO, naturally.

Right now I see new offerings from Gigabyte and Asus. I am hoping to get a mATX offering with all the basic trimmings, such as SATA6 and USB3, and Crossfire/SLI support. I don't need more than the basics and want it in a small envelope.
 
According to Anand, you can't overclock in the H67 boards so you will want to find a matx p67 if that matters to you. I prefer Gigabyte myself, but Asus has its devotees as well. I think either one would treat you well.
 
Just a quick side tangeant... which mobo manufacturers would you recommend for a paired mobo with these CPUs? I've had problems with some lately... so I'm cautious. I think it's important to pair the CPU with a good MOBO, naturally.

Right now I see new offerings from Gigabyte and Asus. I am hoping to get a mATX offering with all the basic trimmings, such as SATA6 and USB3, and Crossfire/SLI support. I don't need more than the basics and want it in a small envelope.

http://hardforum.com/showthread.php?t=1572823
 
Just a quick side tangeant... which mobo manufacturers would you recommend for a paired mobo with these CPUs? I've had problems with some lately... so I'm cautious. I think it's important to pair the CPU with a good MOBO, naturally.

Right now I see new offerings from Gigabyte and Asus. I am hoping to get a mATX offering with all the basic trimmings, such as SATA6 and USB3, and Crossfire/SLI support. I don't need more than the basics and want it in a small envelope.

i ordered the ASUS P8P67 PRO mobo on blind faith. $181 shipped?....SOLD :)
 
I want to see reviews on the P8P67 mATX boards, since my rad will get in the way of anything past 4 slots anyway I may as well get a matx board. I'm prepared to pay for a UD7 though so I want to make sure this matx board can oc just as well.
 
Yeah makes me almost regret getting the Intel RetailEdge promo for the Core i7 970...

This CPU destroys it in gaming...at a much lower power envelope. :rolleyes:
 
I want to see reviews on the P8P67 mATX boards, since my rad will get in the way of anything past 4 slots anyway I may as well get a matx board. I'm prepared to pay for a UD7 though so I want to make sure this matx board can oc just as well.

Hmm, qood question... do mATX typically skimp in overclocking potential, stability, or other areas that would deter from buying them beyond the listed feature-set?
 
Hmm, qood question... do mATX typically skimp in overclocking potential, stability, or other areas that would deter from buying them beyond the listed feature-set?

the consensus around the web (FWIW) is that the board plays a much less significant role in the overclocking potential, since OC'in is achieved almost entirely through the multiplier and not the conventional method of multiplier/FSB combination. time will tell if this is the case.
 
Back
Top