Intel i7-6700K Skylake CPU Benchmarks Appear

[Opcode]

Yeah, it's the beginning. It's a big step compared to just not being relevant at all.

 

Skylake is an architecture that will span all the way from high-power servers, workstations and gaming desktops over laptops to tablets (or at least convertibles). The needs of the mobile space dictates how the architecture is designed, to the detriment of desktop enthusiasts in particular (servers are actually fine with sacrificing clocks and IPC to some extent, since driving down power draw yields power efficiency gains that are useful there).

You won't get Skylake into phones and tablets. Airmont is the next step for Intel in that department (14nm die shrunk Silvermont).

[patrickjp93]

Carrizo (15w) benches nearly twice as fast as HD 6000 in 3DMark11 which is Intel's highest skew in that power envelope. They do have a SIMD advantage although power consumption is something that's hard to not take into consideration. On 28nm compared to Intel's 14nm FinFET. Carrizo on 14nm FF would just be ridiculously good in the laptop ecosystem. However AMD still managed to take the A10-7850k and fit all of its performance into a 15w envelope on 28nm.

 

Scoring design wins is only the beginning of Intel having any success in the mobile market (lost over $4.2 billion last year in it). I wasn't referring to their low power mobile footprint anyhow (Skylake is a laptop/desktop/server architecture).

Yeah, no. The highest HD 6000 benchmark is within 22% of Carrizo for 3DMark 11. You're quoting very old figures from junk laptops.

[Opcode]

Yeah, no. The highest HD 6000 benchmark is within 22% of Carrizo for 3DMark 11. You're quoting very old figures from junk laptops.

Can you link the benchmarks you're using?

[Sakkura]

You won't get Skylake into phones and tablets. Airmont is the next step for Intel in that department (14nm die shrunk Silvermont).

Probably not phones, but tablets should be possible when Broadwell is basically there already. Otherwise Cannonlake will handle that.

[patrickjp93]

Can you link the benchmarks you're using?

http://www.3dmark.com/search#/?mode=advanced&url=/proxycon/ajax/search/cpu/3dm11/E/1945/500000?minScore=0&cpuName=Intel%20Core%20i7-5650U%20Processor

 

That puts it within 20-25% the score of the FX8800P across the board until higher resolution graphics where, admittedly, it falls apart.

[Opcode]

http://www.3dmark.com/search#/?mode=advanced&url=/proxycon/ajax/search/cpu/3dm11/E/1945/500000?minScore=0&cpuName=Intel%20Core%20i7-5650U%20Processor

 

That puts it within 20-25% the score of the FX8800P across the board until higher resolution graphics where, admittedly, it falls apart.

The problem is them are 3DMark 11 Entry numbers being compared to 3DMark 11 Performance numbers for the FX-8800P.

 

http://www.notebookcheck.net/Intel-HD-Graphics-6000.125588.0.html

 

HD 6000 3DMark 11 Performance = 1448

[Techhog]

I would like to point out that the xeon has a lower clockspeed but the same score.

... And twice as many cores. And before you sat that doesn't matter, it beat the 4790K. Clearly having less load on that single core had an effect.

[NumLock21]

Not sure how they got those scores, in CB15 multi-core test. their Xeon E5 2620 v3 got 894. Same CPU here, max I can get, no overclocking is 796, that's a difference of 98 points. Best thing would be to wait for more reputable sites for official Skylake benchmarks. Anyone with the exact same cpus on those list can try it for themselves to see if the score matches or if it's even near it, at stock clocks.

[patrickjp93]

The problem is them are 3DMark 11 Entry numbers being compared to 3DMark 11 Performance numbers for the FX-8800P.

 

http://www.notebookcheck.net/Intel-HD-Graphics-6000.125588.0.html

 

HD 6000 3DMark 11 Performance = 1448

Entry to entry mon frere.

[patrickjp93]

Not sure how they got those scores, in CB15 multi-core test. their Xeon E5 2620 v3 got 894. Same CPU here, max I can get, no overclocking is 796, that's a difference of 98 points. Best thing would be to wait for more reputable sites for official Skylake benchmarks. Anyone with the exact same cpus on those list can try it for themselves to see if the score matches or if it's even near it, at stock clocks.

Better RAM and possibly a BCLK or cache overclock.

[NumLock21]

Better RAM and possibly a BCLK or cache overclock.

that might be possible, going to try that out to see if my score will get closer to what they got. But I only got 2 sticks of ram in dual channel, they're probably running with 4 sticks in quad channel.

[Opcode]

Entry to entry mon frere.

i7-5650U  — 3DMark 11 Entry (1.0) = E2549

i7-5650U  — 3DMark 11 Entry (1.0) = E2517

i5-5250U  — 3DMark 11 Performance (1.0) = P1448

FX-8800P — 3DMark 11 Performance (1.0) = P2645

 

I'm failing to see where you're getting equivalent numbers out of different benchmarks.

[DeViLzzz]

They can keep bringing out new cpus and new games and until I get through much of my current back log of games I won't be needing anything newer than my i5 2500K.  

[patrickjp93]

 

i7-5650U  — 3DMark 11 Entry (1.0) = E2549

i7-5650U  — 3DMark 11 Entry (1.0) = E2517

i5-5250U  — 3DMark 11 Performance (1.0) = P1448

FX-8800P — 3DMark 11 Performance (1.0) = P2645

 

I'm failing to see where you're getting equivalent numbers out of different benchmarks.

 

1) You chose the bottom of the barrel of the HD 6000s with the lowest clock rate

2) Look at the entry scores for FX-8800P.

[Opcode]

1) You chose the bottom of the barrel of the HD 6000s with the lowest clock rate

2) Look at the entry scores for FX-8800P.

1) Lowest clock rate? They are benched at 950 MHz which is the reference clocks maximum boost level.

2) If you can link them I personally have yet to see any.

 

Even if they exist they don't justify anything as no one plays at 800x600. Especially with pretty much all laptops being above HD resolution (1366x768).

[patrickjp93]

1) Lowest clock rate? They are benched at 950 MHz which is the reference clocks maximum boost level.

2) If you can link them I personally have yet to see any.

 

Even if they exist they don't justify anything as no one plays at 800x600. Especially with pretty much all laptops being above HD resolution (1366x768).

max boost level is 1000MHz. Please more carefully read the charts. And actually most of the laptops carrying the 5250 throttle like mad, so yes, let's stick with the 5650 scores, shall we?

[Godlygamer23]

Same old same old, not much has changed since sandy bridge, and I don't expect much to change until AMD gets their act together.... if ever.

Or someone else challenges Intel.

[patrickjp93]

Or someone else challenges Intel.

That's one reason Intel's holding AVX 512 from the consumer chips. It's actually a good thing to have in the HPC space to better chip away at the Power 8 and Power 9 systems which have extremely dense performance for business software and databases, though they're hellishly expensive. If Intel can start winning in the SAP benchmark suite, IBM will be ruined (again) and Oracle will be backed into a corner.

[Opcode]

max boost level is 1000MHz. Please more carefully read the charts. And actually most of the laptops carrying the 5250 throttle like mad, so yes, let's stick with the 5650 scores, shall we?

I would actually tell you to be very careful reading the charts.

 

 

The 5650U are E (Entry) scores compared to P (Performance) scores that we have of Carrizo. FYI if you don't know there is a huge difference between the two benchmarks.

[patrickjp93]

I would actually tell you to be very careful reading the charts.

 

 

 

The 5650U are E (Entry) scores compared to P (Performance) scores that we have of Carrizo. FYI if you don't know there is a huge difference between the two benchmarks.

We have the E scores for the 5650 as well. 2549

[Opcode]

We have the P scores for the 5650 as well.

All you've linked to is Entry scores. What's the point in doing that unless you're very unfamiliar with 3DMark11?

 

Here I did the work for you, the #1 performance score in 3DMark11 database.

 

5650U = P1615 (Graphics Score 1479)

8800P = P2645 (Graphics Score 2700)

 

Still quite a big difference.

[patrickjp93]

All you've linked to is Entry scores. What's the point in doing that unless you're very unfamiliar with 3DMark11?

 

Here I did the work for you, the #1 performance score in 3DMark11 database.

 

 

5650U = P1615

8800P = P2645

Sorry, I meant E. We have the E scores for 5650 and FX 8800P. 5650U tops out at 2549, and FX 8800P tops out at 3102.

[Opcode]

Sorry, I meant E. We have the E scores for 5650 and FX 8800P. 5650U tops out at 2549, and FX 8800P tops out at 3102.

You have the link for the FX-8800P E scores? I haven't seen them yet.

[patrickjp93]

You have the link for the FX-8800P E scores? I haven't seen them yet.

It only presents me the ranges, with no "approved" scores since 3DMark takes forever to get new GPUs and drivers "approved." Just enter it in the search bar and choose entry.

[Opcode]

It only presents me the ranges, with no "approved" scores since 3DMark takes forever to get new GPUs and drivers "approved." Just enter it in the search bar and choose entry.

Did, with 0 results. It's easier if you just link the highest result. The search engine doesn't know what these chips are yet.