A Broadwell Retrospective Review in 2020: Is eDRAM Still Worth It?
by Dr. Ian Cutress on November 2, 2020 11:00 AM ESTBroadwell with eDRAM: Still Has Gaming Legs
As we crossover into the 2020s era, we now have more memory bandwidth from DRAM than a processor in 2015. Intel's Broadwell processors were advertised as having 128 megabytes of 'eDRAM', which enabled 50 GiB/s of bidirectional bandwidth at a lower latency of main memory, which ran only at 25.6 GiB/s. Modern processors have access to DDR4-3200, which is 51.2 GiB/s, and future processors are looking at 65 GiB/s or higher.
At this time, it is perhaps poignant to take a step back and understand the beauty of having 128 MiB of dedicated silicon for a singular task.
Intel’s eDRAM enabled Broadwell processors accelerated a significant number of memory bandwidth and memory latency workloads, in particular gaming. What eDRAM has enabled in our testing, even if we bypass the now antiquated CPU performance, is surprisingly good gaming performance. Most of our CPU gaming tests are designed to enable a CPU-limited scenario, which is exactly where Broadwell can play best. Our final CPU gaming test is a 1080p Max scenario where the CPU matters less, but there still appears to be good benefits from having an on-die DRAM and that much lower latency all the way out to 128 MiB.
There have always been questions around exactly what 128 MiB of eDRAM cost Intel to produce and supply to a generation of processors. At launch, Intel priced the eDRAM versions of 14 nm Broadwell processors as +$60 above the non-eDRAM versions of 22 nm Haswell equivalents. There are arguments to say that it cost Intel directly somewhere south of $10 per processor to build and enable, but Intel couldn’t charge that low, based on market segmentation. Remember, that eDRAM was built on a mature 22 nm SoC process at the time.
As we move into an era where AMD is showcasing its new ‘double’ 32 MiB L3 cache on Zen 3 as a key part of their improved gaming performance, we already had 128 MiB of gaming acceleration in 2015. It was enabled through a very specific piece of hardware built into the chip. If we could do it in 2015, why can’t we do it in 2020?
What about HBM-enabled eDRAM for 2021?
Fast forward to 2020, and we now have mature 14 nm and 7 nm processes, as well as a cavalcade of packaging and eDRAM opportunities. We might consider that adding 1-2 GiB of eDRAM to a package could be done with high bandwidth connectivity, using either Intel’s embedded multi-die technology or TSMC’s 3DFabric technology.
If we did that today, it could arguably be just as complex as what it was to add 128 MiB back in 2015. We now have extensive EDA and packaging tools to deal with chiplet designs and multi-die environments.
So consider, at a time where high performance consumer processors are in the realm of $300 up to $500-$800, would customers consider paying +$60 more for a modern high-end processor with 2 gigabytes of intermediate L4 cache? It would extend AMD’s idea of high-performance gaming cache well beyond the 32 MiB of Zen 3, or perhaps give Intel a different dynamic to its future processor portfolio.
As we move into more a chiplet enabled environment, some of those chiplets could be an extra cache layer. However, to put some of this into perspective.
- Intel's Broadwell's 128 MiB of eDRAM was built (is still built) on Intel's 22nm IO process and used 77 mm2 of die area.
- AMD's new RX 6000 GPUs use '128 MiB' of 7nm Infinity Cache SRAM. At an estimated 6.4 billion transistors, or 24% of the 26.8 billion transistors and ~510-530mm2 die, this cache requires a substantial amount of die area, even on 7nm.
This would suggest that in order for future products to integrate large amounts of cache or eDRAM, then layered solutions will need to be required. This will require large investment in design and packaging, especially thermal control.
Many thanks to Dylan522p for some minor updates on die size and pointing out that the same 22nm eDRAM chip is still in use today with Apple's 2020 base Macbook Pro 13.
120 Comments
View All Comments
Jorgp2 - Monday, November 2, 2020 - link
Why didn't he buy it then?There was even a many lake refresh
Oxford Guy - Wednesday, November 4, 2020 - link
He didn’t buy a desktop high TDP Skylake with EDRAM because it was never produced. Intel decided to sell less for more, which it could so safely since our capitalist system is mainly about a near-total lack of competition far more often (in tech at least) than anything else. Read for comprehension.Oxford Guy - Monday, November 2, 2020 - link
So, the take-away here is that Intel was heavily sandbagging — not bothering to take advantage of the benefit EDRAM provides (for gaming especially).$10 worth of parts and gamers were expected for fork over big money for less performance.
Hooray for lack of competition.
Nictron - Tuesday, November 3, 2020 - link
My i7 5775c died last week after 4.5 years of service at OC 4.1-4.2 Ghz. Now seeing this review I am quite sad as it could’ve gone a bit more.Upgraded to R5 3600XT for now and can always go 5000 series in future on the X570 platform.
Hope competition stays strong!
Oxford Guy - Wednesday, November 4, 2020 - link
You can hope or you can look at the facts. It hasn’t been strong. That’s why Intel was able to sandbag so extremely.alufan - Tuesday, November 3, 2020 - link
The Intel skew on this site is getting silly its becoming an Intel promo machine!You Benchmark but like many Laptop providers you hamstring the AMD CPU with the worst and slowest components, we all know Ryzen CPUs work best with fast RAM in fact you have stated so yourselves in the past on this very site yet you now choose to test the Ryzen option with CL22 bargain basement RAM..... makes me wonder, how much did Intel pay for this review of a 5 year old CPU just to keep the Blue option at the top of the page, Anandtech is a shameful parody of a neutral review site and frankly the owners and editors have exchanged integrity for well whatever you want to call it, shame on you
Ian Cutress - Tuesday, November 3, 2020 - link
I test at JEDEC speeds. For Ryzen that's DDR4-3200, and JEDEC subtimings are CL22. If you want to complain, complain to JEDEC to ask for something faster, or ask companies to validate memory controllers beyond JEDEC standards. Otherwise it's overclocking, and if we're overclocking, then who cares about core frequency or power anyway.https://www.youtube.com/watch?v=jQe5j7xIcog - I even did a video on it.
I do tons of AMD coverage. Year in year out interviews of CEO and CTO of AMD, but no equivalent of Intel. Deep dives into every major architecture, with analysis of instruction performance as well as cache hierarchy. Reviews of almost every single Ryzen product, as we're sampled most of them. If we were that big of an Intel shill, why does AMD supply us what we ask for?
alufan - Tuesday, November 3, 2020 - link
Ahh so the Ram kit you used for the Intel test just happened to fall into the slots then with its CL16 is that JEDEC standard?I dont want any particular preference shown to either Brand I will admit however to not liking Intel due to former poor experience but I think its important when charts and such are published that a level field is used, because at some point somebody may well use those published figures as a illustration of one products superiority over another, Intel is very good at doing just that, and I will cite the last Thread ripper release for that one just so they could have a set of tables showing the Intel product on the Top even if it was for a few hours.
qwertymac93 - Tuesday, November 3, 2020 - link
Where did you find the latency settings the tests were performed at? I didn't see the latencies mentioned in the test setup page.alufan - Tuesday, November 3, 2020 - link
"Where did you find the latency settings the tests were performed at? I didn't see the latencies mentioned in the test setup page."look up the parts on the makers website