Intel BTFO
Images are sometimes not shown due to bandwidth/network limitations. Refreshing the page usually helps.
You are currently reading a thread in /g/ - Technology
You are currently reading a thread in /g/ - Technology
Intel BTFO
is AMD goin to release three cpu platforms at once??
>>54669652
*significant performance leap relative to their other products.
Wew. something 50%+ better than shit is still garbage.
>>54669673
>using isolated nouns as adjectives
I don't want plebeians like you to exist. Please kill yourself.
>>54669652
that second graph tho
Lines on the two graphs dont match up...
>>54669858
That's fine, they are different graphs. The right one is useless though
>>54669652
>Bristol Ridge ahead of schedule
Awesome. I want to ditch my Kaveri chip, can't wait to see what Bristol ridge is like. It also promises an upgrade pathway since the motherboards will be unified.
>>54669652
So with Zen, are they planning to put a 4 core chip against the Intel "gamer" SKUs? Or are they gonna bank on 8 cores being used better with Vulkan/DX 12?
>>54670366
A very good question, especially considering that there will be two threads per core. Will we be getting 16 thread beasts at i7 prices? Will 8 thread CPUs be the new mid range?
>>54670366
>>54670418
or they could do both
damn genius
>>54669652
>multi threaded compute
I don't care, they need significant IPC changes before they can become competitive with Intel
>>54670462
if they reach sandybride-e performance by now, that would be good enough for most people.
>>54670462
IPC for zen is supposedly able to compete at around a haswell level. I think we'll see it become decent, but still fall short. It's definitely going to depend on the pricepoint.
>>54670462
AMD said in their investor conference that they were exceeding their 40% IPC gain goal already
>>54670551
>>54670556
Sandy bridge performance wouldn't cut it for me, I already have that, and have had it for a long time.
Now if they can release a 6 or 8 core with Hyperthreading for a decent price, then I might be interested even if IPC would only be at sandy bridge level.
More likely the first ZEN cpu's will be expensive though since we are talking 6 and 8 core ones
>>54670462
http://www.itwire.com/it-industry-news/development/72853-amd-zen-will-favourably-compete-with-intel-skylake.html
>>54670606
The 8 core was said to be aimed at "under 400USD", attempting to hit skylake-e with price
>>54670614
>>54670638
interesting read
400 bucks for a haswell IPC level 8 core with HT would be insane value for money
THANK YOU BASED SHIT WRECKER
How many years will they release those graphs? Maybe they should release the actual CPU instead?
>>54670782
Engineering samples are being sent to "select customers" in Q2
>>54669881
>The right one is useless though
missing numbers added.
>>54670988
Your ideas intrigue me and I wish to subscribe to your newsletter.
>>54669652
From what we've seen Bristol Ridge would be pretty good for really small mini-ITX builds
Twice as fast BD?
Wasn't it 40%?
What the fuck is Bristol Ridge? Isn't that just Carizo? Just wait for Raven Ridge, aka actual Zen cores in an APU
>>54672482
Yep, its just Carrizo for desktop with DDR4.
>>54672482
Bristol Ridge itself isn't so exciting but the new socket/platform is if you're a zen waitfag
Hopefully Zen makes ultrabooks and its ilk actually useful and gets rid of 2 cores in $1000+ laptops.
I fell for the 8350 because it was on sale, but it wasn't so bad. This one will probably be in line with that expectation and hopefully have the same power in a single core that an i5 has
>>54672451
The 40% wasn't in relation to bulldozer/vishera - it was in relation to their latest APU range (which basically murder bulldozer in some areas).
Believe it or not AMD has made some decent IPC gains in their APU chips but (for many complicated reasons) have not released a high performance chip since vishera.
>>54669711
Except it makes more sense when read as a metaphor than as adjectives. I don't want like plebs like you in this thread.
>>54672617
>but (for many complicated reasons) have not released a high performance chip since vishera.
It's not a complicated reason. AMD just didn't want to invest money into a Steamroller FX chip knowing that it would tank.
>>54672617
>t was in relation to their latest APU range (which basically murder bulldozer in some areas).
No real product (CPU or APU) does. Sure, in theory Kaveri and Carrizon do 'murder' bdver2 at some tasks, but since APU's based on those are gimped in cakes and whatnot, they fall flat.
>>54672617
Steamroller is a very small uplift over Piledriver. L2 latency dramatically increased which hurt performance in everything and that destroyed what performance was gained from most of the architectural tweaks.
Excavator on average is a bit faster than Steamroller, and does really well in certain benches, but can also be slightly slower in some. Over all its not even directly competing against Steamroller in anything but perf/watt.
Look up the Carrizo reviews done by the Stilt.
AMD couldn't salvage any performance out of the Bulldozer family so they just turned it into a low power arch to replace the retired Cat cores.
>>54672708
>the Stilt
The Finnish AMD CPU enthusiast.
>>54672791
He does know his stuff though. See also: Buildzoid for AMD gpu murdering.
