So is ryzen 2 going to be as good as intel for games...

So is ryzen 2 going to be as good as intel for games? What is the reason that ryzen sucks at games compared to intel atm?

Attached: 1532711557073.jpg (1200x600, 169K)

Other urls found in this thread:

semiaccurate.com/2018/08/02/intel-guts-10nm-to-get-it-out-the-door/
ht4u.net/reviews/2013/intel_core_i7_4770_4670_haswell_cpus_test/index37.php
twitter.com/SFWRedditVideos

amd still can't into single thread performance which is what most things care about.

TSMC/GloFo's 7nm is on par with Intel's 14nm++.

>intel
>good

ryzen 2 is the ryzen 2xxx series (which is zen+) and is already on the market
I believe you meant zen 2 which will be the ryzen 3xxx series

Ryzen is not shit but games are
Games like Arma 3 that just run poor on anything not 5ghz should be purged
How can you critizie prowsers for runnig poor but support games at pajeet level
>if my predictions are right

>So is ryzen 2 going to be as good as intel for games?
Zen 2, but yes from a technological point they should see better performance than any Intel chip then(Unless Intel pushes also out some 7mm solution till then, but no one expects that).

>What is the reason that ryzen sucks at games compared to intel atm?
Lower single core clockspeed, which is what most games still rely on. Unlike Intels consumer chips, Zen isn't focused on a low core high power design, but rather the other way round. With a focus on infinity fabric is basically a "dumbed down" version of a server chip. The obvious results is what we're seeing in every benchmark: single core performance: Intel wins, multi core performance: AMD wins. Now the consumer can decide what he needs and choose appropriately.

Attached: d0106319_4f37aee46b64c.jpg (743x1051, 100K)

>Unless Intel pushes also out some 7mm solution till then
>7mm solution
Got you senpai

Attached: 1489122339460.png (1070x601, 495K)

Meanwhile, in reality...

Attached: 1535561310892.png (731x918, 65K)

Damn I wish that were me

more reality...
semiaccurate.com/2018/08/02/intel-guts-10nm-to-get-it-out-the-door/

Intel "10"nm is now "12"nm

>muh bideo bames

Some games give the 4770K 31% more performance over the 3770K

ht4u.net/reviews/2013/intel_core_i7_4770_4670_haswell_cpus_test/index37.php
Is this an application of AVX2?

There is nothing wrong with ryzen in gaymen fucking dumb /v/ faggot

Thats only half of it. This image is showing the metrics of the Connected Poly Pitch, and the Min Metal Pitch.
TSMC has smaller SRAM cell size than intel, GloFo and Samsung do as well. This is huge because for years intel had the densest caches in the industry.
TSMC as well is leading intel in total logic cell area by having lower Z height.

Its very likely that intel is no longer using 54nm and 44nm for their MMP and CPP, they've probably scaled back to improve yields. They will never officially confirm it, but everyone knows it, fab workers talk, people like Charlie at S|A pays for insider info.
Their process could be around 60nm and 48nm or there abouts, making their cell size much larger than previously stated, making their area scaling worse, negating the only big key feature they aggressively pushed for from the very beginning.

Zen 2 will likely be even better in games than what intel currently has.
1.) Architectural improvements will increase the IPC/MIPS(without clocks) and since AMD's Current IPC is within 3-5% of Intel, we will likely see a 5-10% increase over Intel's IPC
2.) Currently Ryzen doesn't have proper AVX2 Support, (Which is used widely in games). As I understand it, it is actually just one AVX section per half-ccx that can work in tandem with eachother and also have support for specifically AVX2 instructions while not actually being 256 wide. AMD is almost for sure going to fix this with Zen 2, and will likely have the same configuration I mentioned, but each half of the CCX will have proper AVX2 that work together to fake AVX-512.
3.) Zen 2 is going to be 7nm, This will allow them to cram more transistors into a same-size die than even intel's 10nm which isn't out. This will likely lead to larger Cache, and the development of current/new architecture components. On top of this, the clock speed will be higher; if I remember correctly they are using TSMC's 7nm-LP which will target 5Ghz frequency (GF's 14nm Targeted 3Ghz). So AMD will likely have their processors clocked around 5Ghz, although it is worth mentioning that this is an immature process and may not clock that high for a few years. On top of this, Infinity Fabric will likely clock higher as well, which leads me to:

>cont.

Did they solved the Micro housefire issue that they had with cobalt liners?

4.) Infinity Fabric is the current main bottleneck in terms of the Ryzen Architecture (except for maybe AVX support), for data to get from one Half of a CCX to another, it will take around 60ns to travel. This limits how quickly information can be processed when transferred between these modules. On top of this, current operating system and application schedulers Don't really know how to deal with this structure of non-uniform Memory Access (NUMA). This leads to decreased FPS in games as information travels over the infinity fabric. With 7nm, the infinity fabric will likely be able to clock higher than it had previously, cutting down on the latency penalty.

This is all assuming that AMD doesn't increase the core count absurdly high like I said previously. At 8 Cores, the architecture is already mildly starved of Memory Bandwidth with only Dual-Channel DDR4. This problem will only be exacerbated if AMD doubles the core count.

I highly doubt that they will dare to spit towards the sky. I mean, Ryzen is bretty gud, its taking marketshare in the server market, etc. But AMD still needs foothold to mantain its momentum.

The IF data rates and ratios aren't fixed, they can all be tweaked. AMD will lower signal latency as process node allows.

Lolno, they have to reengineer the node with different material contacts. 10nm will come out, just not with cobalt for high volume production.

Zen 2 isn't Ryzen 2