Intel PR claims RKL-S offers superior storage performance

[HEXUS]

And, a Core i5-11400 Geekbench leak shows it has 24 per cent better 1T perf than its predecessor.

Read more.

[Tabbykatze]

You forgot to note about the "Quick Storage Benchmark" that it specifically says it's to "measure the performance of small system drives from traditional spinning drives (HDDs) at the low end and entry-level PCI Express SSDs at the high end".

So a PCIe 4.0 high performance SSD is likely being CPU constrained by the method this benchmark utilises to analyse the performance of the SSD.

That, and it's Ryan Shrout so anything he puts up should be criticised from end to end quite harshly due to his history with AMD benchmarking.

[Corky34]

From the people who say we shouldn't be using benchmarks but 'real world' performance comes, a benchmark.

[LSG501]

So that would be a full, non trimmed, drive on the AMD system and an empty one on the Intel then....yes I'm being cynical, I'll wait for unbiased reviews.

There are so many ways to 'nerf' the performance of an ssd drive in benchmarks I wouldn't take one from a company who is 'on the back foot' at the moment too seriously.

[trillo_del_diavolo]

Yeah it will remain in the shelves for a long time...

[philehidiot]

It's a relief to see that everyone is sceptical about anything from Intel, no matter how minor.

Bearing in mind this really is scraping the barrel. Most consumers have more storage performance than they know what to do with. For me, for general computing tasks, storage has not been a bottleneck since my first X-25M SSD. For silly tasks which should really necessitate a Threadripper and semi-pro storage that I'm trying to do on a consumer grade system, well yes it bottlenecks. But usually for short bursts that cost me about a minute a week averaged over a year.

[edmundhonda]

the Intel Chief Performance Strategist had to reply to clear up this query, asserting that both systems used the 1TB Samsung 980 Pro installed on an x16 PCI riser card.

Unnatural test setup.

It's a relief to see that everyone is sceptical about anything from Intel, no matter how minor.

Besides the obvious detail of a company benchmarking their own product, they've deliberately obscured the actual numbers. It's a public test that should produce results within a narrow window, but Intel have explicitly chosen to present the outcome as '11% faster' so there's no way to anyone to crosscheck the baseline result. Pretty typical!

[ohmaheid]

Anything emanating from Ryan Shrout should be treated with deep scepticism.

[philehidiot]

Anything emanating from Ryan Shrout should be treated with deep scepticism.

Unless it's a turd. Then it's probably actually a turd.

[Raven74]

Anything emanating from Ryan Shrout should be treated with deep scepticism.

Unless it's a turd. Then it's probably actually a turd.

Nope, if it says so then it would be the biggest turd, dont believe it is just a turd.

[Yoyoyo69]

Not to defend AMD or diss Intel, but it's not exactly a huge difference (but granted a positive one "if" it were real).

Like others I'm extremely doubtful if ANYTHING Intel push out, even if they could provide a large amount of proof.

They are so desperate, they have to try and brag and promote tiny positives, which usually end up being proven either fake or very highly skewed.

I'm torn, I want competition, but I'd much rather Intel weren't rewarded for being the skum of the Earth.

[Corky34]

Not to defend AMD or diss Intel, but it's not exactly a huge difference (but granted a positive one "if" it were real).

To be fair to them it's probably mainly down to Intel's better single core clock speed/performance (*on Windows) vs the AMD chip they tested against, a 5950X boosts to 4.9 on single core whereas (if rumours are correct) the Intel chip has a 400Mhz advantage.

*I think Windows only uses a single thread for file transfers IIRC.

[gagaga]

Amazing that you can get a significant boost to IPC when you haven't updated your architecure for 5 years.