Report: The Intel Core i7, in practice (page 12: Conclusion)

Conclusion

At a time when the Core 2 is already so dominant, has Intel succeeded in doing even better with the Core i7? After having studied Nehalem architecture in an earlier report, we can confirm that its implementation into the Core i7 has been a success.

Indeed, the various optimisations that have allowed Core architecture to evolve into Nehalem are bearing fruit and the performance gains are significant, even if a little disparate. The essence of the benefits of this processor lie in multithreaded applications, which are for the most part those we use in our test protocol. Largely due to SMT, gains of between 20 and 40% are achieved on the Core 2 at equal frequency, which is to say the least remarkable given the already impressive performances on the Core 2!

Gamers, who benefitted greatly from the Core 2, won’t see much improvement on the Core i7. Game performance gains are low and there’s even sometimes a slight fall in performance at the same frequency, something we didn’t include in our test protocol. With lower consumption and at a much more affordable platform cost (motherborad at €100 or €150, DDR2…), the Core 2, whether dual or quad, is still without a doubt the best choice for these users, except if you absolutely want a platform capable of SLI, like CrossFire.

So, the Core i7 can set out its stall as the ultimate workstation processor, but not so much a CPU for the general public. Intel is well aware of this and the X58/LGA1366 platform is more of an advertisement than a product destined for mass volumes. Although peformance nuts and pros are likely to jump a them, others will no doubt wait for the LGA1160 platform, expected in mid 2009.