Tuesday, July 5th 2011
Intel Ivy Bridge Dual-Core Put Through Clock-to-Clock Benches Against Sandy Bridge
Taiwanese PC enthusiast Coolaler has a new Ivy Bridge LGA1155 dual-core engineering sample to play with, and wasted no time in putting it through some tests. The sample has two cores, four threads with HyperThreading enabled, clock speed of 1.80 GHz, 256 KB L2 cache per core, and 4 MB shared L3 cache. It is running on an Intel P67 chipset-based motherboard with 8 GB of dual-channel DDR3-1600 MHz memory. At 1.80 GHz, it may not be game for absolute performance figures since it's unlikely that Intel will release a chip with that clock speed unless it has unreal performance:clockspeed gains over Sandy Bridge; but it's good enough for clock-to-clock performance comparisons between Ivy Bridge and Sandy Bridge. A Core i5-2400 was clocked at 1.80 GHz with 18x BClk multiplier, and put through a single-threaded benchmark, and a multithreaded one.
The cache and memory benchmark that measures bandwidth and latency of caches and memory was unable to measure bandwidth, but measured some latencies. The L1 cache at 2.2 nanoseconds (ns), and L2 at 2.9 ns. Next, the Ivy Bridge DC, and the Core i5-2400 (@ 1.80 GHz) testbeds were put through CPUMark 99, where Ivy Bridge DC scored 278 points, and Core i5-2400 clocked at 1.80 GHz scored closely followed at 276 points. Moving on to multithreaded performance, the two were put through Cinebench 11.5 64-bit. The Ivy Bridge DC chip scored 1.81 points; while the Core i5-2400 clocked at 1.80 GHz, scored 2.61 points. Coolaler promises more benches.
Source:
Coolaler
The cache and memory benchmark that measures bandwidth and latency of caches and memory was unable to measure bandwidth, but measured some latencies. The L1 cache at 2.2 nanoseconds (ns), and L2 at 2.9 ns. Next, the Ivy Bridge DC, and the Core i5-2400 (@ 1.80 GHz) testbeds were put through CPUMark 99, where Ivy Bridge DC scored 278 points, and Core i5-2400 clocked at 1.80 GHz scored closely followed at 276 points. Moving on to multithreaded performance, the two were put through Cinebench 11.5 64-bit. The Ivy Bridge DC chip scored 1.81 points; while the Core i5-2400 clocked at 1.80 GHz, scored 2.61 points. Coolaler promises more benches.
70 Comments on Intel Ivy Bridge Dual-Core Put Through Clock-to-Clock Benches Against Sandy Bridge
2. It's an ES Dual Core
Wait a few months for some good silicone to come out. Also why would intel produce a processor thats 20% better then a processor they just released 7 months ago?
Hopefully there will be some nice high res shots of ivy although I'm doubtful that the change in transistor will be visible as most die shots are not high enough res, a 500 megapixel shot of ivy bridge would be nice though :laugh:.
Yes there is some great detailed shots available but come on, you know you would look at a 100mp+ die shot if you could :p
I hope Intel really hypes up the new transistor as ivy bridge comes closer to release so there will be more information and maybe even some close up shots to show off the transistors themselves :D although that last bit may be a little too hopeful.
Once pictures get so small i get easily confused as there is so little context, I assumed due to the gaps in what i assumed was the conductive material it was not a circuit, what i was hoping for was a shot like that of the ivy bridge die but have i been a dumb ass and already looked at the shot i wanted? :laugh:
But i so can't wait to find out how these transistors effect the CPU's, there is so much more gate material it would be great if it makes a big difference to power consumption or max speed/overclocking although as it comes with a die shrink i guess it's going to be hard to judge how much comes from the shrink and how much comes from the new transistor.
2. He run Super PI which is single threaded so that Dual core vs Quad argument meens sweet FA...
But the 1000 times density kind of confuses me as it seams too much, i may be far off but would 1000 times denser mean that if they used the new transistor to make a CPU with the same die area as sandy bridge (216 mm2?) it could fit 915 billion transistors in that space instead of 915 million? are these truly 1000 times smaller than the current transistors Intel uses?
I never noticed anything about size difference in the TPU news article on the transistors although that was months ago so i may have just not noticed or forgotten :laugh:
Even single threaded applications can benefit from multiple cores, because the additional cores can carry out processing on background applications and hence allowing the main processor to compute SuperPI. Hence why the Ivy is slightly behind.
I would love to see any gain with a 20% power consumption reduction.