• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Intel Broadwell Thread

Page 112 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
I'm somewhat disillusioned at this point. I see some prebuilt systems with these DT chips, but enthusiasts have clearly been left out in the cold. Need to find a new hobby.
 
If Intel comes with their own version of HBM... no one would need nVIDIA anymore too 😀

Well NVidia doesn't have a HBM card out....only AMD does. It performs about equivalent to a GTX 980 but slightly less than a GTX 980 Ti.

The CrystalWell embedded DRAM (128 MB) also doubles as an L4 Cache for the CPU.

I'm really convinced with this type of performance we might see a console from Intel for the 9th generation - and it could potentially be backwards compatible with PS4 / XBOX One.
 
Better performance and efficiency, plus if I'm building HTPC, I want small & no discreet card..

Yes, if you have an absolute criteria that there should be no discrete graphics card and you still want decent GPU performance and don't care about price, then I can see the Broadwell 5675C being a good option. Maybe the 65 W TDP is a bit high though, I myself would preferred 35 W or less if to be used for HTPC. Anyway, I think that's one of the few cases this CPU makes sense.

Otherwise e.g. a combo like frozentundra123456 mentioned here would be a better option, i.e. a i7-860k + 260x combo for around $175. But that's just one example, there are lots of other options that are better than the 5675C.
 
Last edited:
Core i5 6600K will not have an easy time beating Core i5 5675C in CPU limited games that benefit from eDRAM:

Overall
moyennejeuxx_zpsri5s52kr.jpg


totalwarromeii_zpspikd5kmj.jpg


arma3_zpsdxgf4sy9.jpg


anno2070_zps47tdgyjm.jpg


www.hardware.fr/focus/112/core-i5-5675c-broadwell-cote-cpu-test.html

Definitely a viable upgrade path for LGA1150 users, I just hope availability (and retail prices) improve. Too bad Broadwell-K's launch was so close to Skylake-S, I feel like this chip is underrated. The CPU performance it provides for dGPU gaming alone justifies the MSRP (Intel ARK) price, then there's the best in class Iris Pro iGPU and much better efficiency than Haswell (something many reviewers could testify).

If Intel keeps 65W Skylake 4C+GT4e (64MB eDRAM) clocks high enough next year then this might be a better gaming chip than 91W 4C Kabylake-K (assuming Kabylake packs higher-clocked Skylake cores, not a 14nm version of Cannonlake cores).
 
Last edited:
He was refering to AMD 860K :biggrin:

Ahhh, of course! It wasn't clear from his post. How confusing for Intel and AMD to have the same model number like that. But I was too quick to jump the gun, and Intel's 860 is in fact really old (from 2009). 😳 Anyway, with the AMD 860K the point is still valid though.
 
Last edited:
Well NVidia doesn't have a HBM card out....only AMD does. It performs about equivalent to a GTX 980 but slightly less than a GTX 980 Ti.

The CrystalWell embedded DRAM (128 MB) also doubles as an L4 Cache for the CPU.

I'm really convinced with this type of performance we might see a console from Intel for the 9th generation - and it could potentially be backwards compatible with PS4 / XBOX One.

R390X performs about the same as GTX980, Fury and especially Fury X are way above that performance level.
 
some PCLab.pl tests

energia_fifa15.png


w3_hd.png


power efficiency and performance is really weak compared to the 750 on Witcher 3
http://pclab.pl/art64766-3.html


in some other games, it looks better, but Witcher 3 is playable with a $100 card and not playable with Iris pro 6200; something to consider.
 
Lol! And even Godavari manages to beat Broadwell (with low margin).... with a process 3 generations behind!

If AMD could move Bristol Ridge on 22 nm SOI.... only to make time to squash Intel with their HBM Zen APU on GPU (remember that the GPU is really strong... they need dedicated RAM... I am wondering what happens if that GPU has 128 Mb eDRAM)
 
Lol! And even Godavari manages to beat Broadwell (with low margin).... with a process 3 generations behind!

If AMD could move Bristol Ridge on 22 nm SOI.... only to make time to squash Intel with their HBM Zen APU on GPU (remember that the GPU is really strong... they need dedicated RAM... I am wondering what happens if that GPU has 128 Mb eDRAM)

You might want to check the other tests, Broadwell-K beat Kaveri in all of them. 🙂
They probably need some drivers work in The Witcher 3 and some other titles.
Also better get used to Intel's lead cause all AMD has planned for next year in terms of APUs is another 28nm rehash.

If we're doing selective posting, let me join:
sc2.png
 
Two new reviews guys, Techspot and Guru3D

Techspot's Core i7 5775C Review

- iGPU Gaming

iGaming_01.png


iGaming_05.png


iGaming_03.png


iGaming_02.png


iGaming_04.png


iGaming_06.png


Much better performance in Crysis 3 and Tomb Raider, interesting.

- Clock per Clock Comparison

Clock_06.png


- Power Consumption

Power_02.png


- OC Performance

Overclocking_05.png


Overclocking_06.png


Where the 5775C shines is in its 3D performance, and we hope AMD has been taking note. The Iris Pro 6200 makes the 5775C the fastest socketed processor available when it comes to 3D rendering. AMD’s A10-7870K was simply no match for the i7-5775C in any of the games we tested.

The 128MB of eDRAM which acts a bit like a L4 cache was the key to the 5775C’s amazing 3D performance. That said, we suspect it came in handy on the CPU side of things when running a few of our application tests, where the 5775C was surprisingly fast.

www.techspot.com/review/1028-intel-core-i7-5775c-broadwell
 
Last edited:
Guru3D's Intel Core i7 5775C Broadwell-H Review

index.php


index.php


index.php


index.php


index.php


The GT3e GPU that is embedded alongside the L4 cache chip is powerful opposed to what Intel has been releasing for years now. I mean to date there hasn't been a desktop processor from Intel that was capable beating AMD APUs. Broadwell-H Desktop processors come with a very feisty solution, and if you stick to mid-range quality settings, you can actually play games at 1080 reasonably well. Truth of course remains, everybody that purchases a flag-ship quad-core processor, will likely already have a dedicated graphics card. But yeah the GT3e IGP is good, really good. We also need to separate gaming from the graphics core, realize that it is a multipurpose graphics processor. You'll have no issues with Blu-ray playback, heck it can even deal with 3D TV, post-process your media files and help out with video en/trans and decoding. The GPU is compatible with HDMI 1.4 and Display Port 1.2 next to the regular DSUB and DVI connectors of course. Unfortunately HDMI 2.0 is not supported meaning 60Hz on an Ultra HDTV will max out at 30Hz.

The biggest benefits for Broadwell-H are twofold: impressive integrated graphics performance and a notch gained in power efficiency. Somehow I have the feeling that Broadwell-H for desktop is sort of a initial tryout to see how 14nm is behaving. As a result the performance is good, but relatively clocked low. And then overclocking results are a bit dull. The performance increment overall is close to NIL. So with that said, if you already are on Sandy-Bridge, Ivy-Bridge or Haswell, we would not recommend you to upgrade. See, what you want to wait for is Skylake, which is launching very soon already. That desktop processor series promises new motherboards platforms with new features and a processor that will get you a fair bit more in performance and tweakability as well. In retrospect I do have to mention that the maximum turbo allowance for the 5775C is 3.7 GHz, and it can keep up with say a 4790 really well. So in that respect the per core performance has increased a bit.

www.guru3d.com/articles_pages/core_i7_5775c_processor_review_desktop_broadwell,1.html
 
For now, but Broadwell-K and the new line of Iris Pro desktop chips is here to stay.
If this week's leaked roadmap is correct higher clocked Broadwell-K arrives in Q3. Skylake GT4e should follow next year.
 
The fastest unobtainable mainstream CPU ever produced.

It is available in Germany.


If this week's leaked roadmap is correct higher clocked Broadwell-K arrives in Q3. Skylake GT4e should follow next year.


Looks like you don't understand Intels Roadmap. Intel might look into a higher clocking variant option but there is no definite hint from the >= sign that it comes. In the past we didn't see higher clocking variants even with this sign in a Roadmap.
 
Looks like you don't understand Intels Roadmap. Intel might look into a higher clocking variant option but there is no definite hint from the >= sign that it comes. In the past we didn't see higher clocking variants even with this sign in a Roadmap.

Since Skylake GT4 is H2/2016 they most likely will release higher clocked variants, doesn't matter if it's Q3 or later. I still stand by what I said, availability should improve in the coming months and desktop Iris Pro is here to stay.
 
Well NVidia doesn't have a HBM card out....only AMD does. It performs about equivalent to a GTX 980 but slightly less than a GTX 980 Ti.

The CrystalWell embedded DRAM (128 MB) also doubles as an L4 Cache for the CPU.

I'm really convinced with this type of performance we might see a console from Intel for the 9th generation - and it could potentially be backwards compatible with PS4 / XBOX One.

What you are saying implies that there's not much of a performance gap between a 980 and a 980Ti but there is. 980Ti at stock is faster than 980 at stock by about 30%. 390X performs about on par with 980. Fury/FuryX are at a (heavily) overclocked 980 level. OCed 980Ti is in a league of its own. Overclocking 980Ti gives more of a performance jump then the one you get from stepping up to 980ti from 980.

http://www.3dmark.com/fs/5396318

my results practically matches that of my ex SLI Titans and that is without node-shrink, back in the days such improvements were only possible with both new architecture and die shrink, this time sacrificing DP64 performance and optimization of the architecture sufficed to provide nearly the same jump.
 
Back
Top