hdparm -tT --direct /dev/sda

/dev/sda:
Timing O_DIRECT cached reads: 248 MB in 2.02 seconds = 122.80 MB/sec
Timing O_DIRECT disk reads: 354 MB in 3.01 seconds = 117.55 MB/sec

Same with direct... 112 to 117 is just because the moment of the test... what I'm saying these benchs are just for few seconds... use the HD Tach to you see (I don't have windows do make the test).

Can you share some app for Linux to do long HDD read benchs?

PS. It is the four time I said I was wrong about the BD/HDD thing... did you really read my posts?

I can :) But what do you want to prove? Do you still deny that an hdd is slower than 6xbdd?

"dd if=/dev/zero of=benchfile count=1024 bs=1M conv=fdatasync,notrunc" you can try. And when using ubuntu it has a builtin benchmark tool: gnome-disks.

Bolded: Sorry but are you doing that just to piss me off?

I will try that other.

Oh, wanted to say faster, not slower :)
No I just want to hear that your initial post starting this discussion saying "PS4 didn't have this problem... the DB player is fast than any mechanical HDD... so the devs will never use the HDD to cache the data before sent to RAM." is just wrong and proven wrong. It's no problem to say that it was false.

Please... read my posts again... I'm sure I said four time directly to you "I WAS WRONG ABOUT THAT"... you just continue to ask over and over again to piss me off... there are no other reason.

This could be another huge mistake..  Last gen they didn't reserve enough resources to copy the entire Xbox Live experience and were never able to add cross browser chat. They should reserve a lot right now and once it is known how much Microsoft is reserving then lower it to that level. Sony doesn't want to be in a situation where Microsoft has some really cool features that they can't copy.

The HDD avegare read drop after the first seconds (I guess after the buffer became full)... for real use the average read for HDDs is close 50-60MB/s because they can't sustain these numbers showed by the sites.

The HDD not changed from 2008 to now... it is the same mechanical thing... the companies just added more buffers caches to speed the first second of the reads.

That is what I wanted to demostrated.

I think the problem with the cross voice game chat if more about Sony infra than memory because there are some games with voice chats.

They even reduced the OS memory use from 120MB to 50MB... so it is enough to make the Cross Voice Chat works.

Oh, no offense! Did you stealth-edit your posts? ;) Sorry man, didn't get it. Everything's fine. Goosfraba :)

So how come their infrastructure was changed for Vita but not for PS3?