ASRock Industrial NUCS BOX-1360P/D4 Review: Raptor Lake-P Impresses, plus Surprise ECC
by Ganesh T S on January 29, 2023 9:30 AM EST- Posted in
- Systems
- NUC
- UCFF
- Mini-PC
- ASRock Industrial
- Raptor Lake-P
GPU Performance: Synthetic Benchmarks
Intel did not make significant changes in the integrated GPU when moving from Alder Lake to Raptor Lake. Process maturity has allowed it to clock the iGPU a bit higher, but the number of EUs remains the same as in the previous generation. GPU performance evaluation typically involved gaming workloads, and for select PCs, GPU compute. Prior to that, we wanted to take a look at the capabilities of the iGPU in the Core i7-1360P. Unfortunately, GPU-Z doesn't yet recognize the 'new' GPU, but HWiNFO has more helpful information.
We have seen earlier that the performance of the Intel Iris Xe Graphics is miles ahead of previous iGPUs from both Intel and AMD. The benchmarks processed on the NUCS BOX-1360P/D4 back up that aspect.
GFXBench
The DirectX 12-based GFXBench tests from Kishonti are cross-platform, and available all the way down to smartphones. As such, they are not very taxing for discrete GPUs and modern integrated GPUs. We processed the offscreen versions of the 'Aztec Ruins' benchmark.
Raptor Lake-P's higher iGPU clocks enable it to come out on top, but enabling ECC makes the performance suffer badly.
UL 3DMark
Four different workload sets were processed in 3DMark - Fire Strike, Time Spy, Night Raid, and Wild Life.
3DMark Fire Strike
The Fire Strike benchmark has three workloads. The base version is meant for high-performance gaming PCs. It uses DirectX 11 (feature level 11) to render frames at 1920 x 1080. The Extreme version targets 1440p gaming requirements, while the Ultra version targets 4K gaming system, and renders at 3840 x 2160. The graph below presents the overall score for the Fire Strike Extreme and Fire Strike Ultra benchmark across all the systems that are being compared.
| UL 3DMark - Fire Strike Workloads | |||
3DMark Time Spy
The Time Spy workload has two levels with different complexities. Both use DirectX 12 (feature level 11). However, the plain version targets high-performance gaming PCs with a 2560 x 1440 render resolution, while the Extreme version renders at 3840 x 2160 resolution. The graphs below present both numbers for all the systems that are being compared in this review.
| UL 3DMark - Time Spy Workloads | |||
3DMark Wild Life
The Wild Life workload was initially introduced as a cross-platform GPU benchmark in 2020. It renders at a 2560 x 1440 resolution using Vulkan 1.1 APIs on Windows. It is a relatively short-running test, reflective of mobile GPU usage. In mid-2021, UL released the Wild Life Extreme workload that was a more demanding version that renders at 3840 x 2160 and runs for a much longer duration reflective of typical desktop gaming usage.
| UL 3DMark - Wild Life Workloads | |||
3DMark Night Raid
The Night Raid workload is a DirectX 12 benchmark test. It is less demanding than Time Spy, and is optimized for integrated graphics. The graph below presents the overall score in this workload for different system configurations.
The Wall Street Canyon NUC and the NUCS BOX-1360P/D4 are pretty much neck-to-neck in the 3D Mark workloads. In these benchmarks that run relatively longer than GFXBench, the PL1 also starts coming into the picture. The Wall Street Canyon NUC has an edge in that aspect. Other than that, it is no surprise that the pattern of external memory accesses generated in the 3D Mark workloads is detrimental to performance when ECC is enabled.
Facebook
Twitter
RSS
30 Comments
View All Comments
drajitshnew - Sunday, January 29, 2023 - link
The in band ECC is an absolutely brilliant idea for systems with 64 GB or more. It is unfortunate that windows does not support it.Samus - Sunday, January 29, 2023 - link
My understanding is this doesn't need support at the software level. This is still "hardware ECC" and OS-independent.Samus - Sunday, January 29, 2023 - link
Oh, I see what you are saying. About how Windows will handle an error. In AT's memtest run the test triggered a stop interrupt presumably as it didn't know how to handle the error. I see what you are getting at with Windows.bernstein - Monday, January 30, 2023 - link
it's more likely, that chrome mandates ecc support, while with windows intel pushes ecc as $$$ featuresjkpublic@gmail.com - Monday, February 13, 2023 - link
This competes with laptops. Please expand on why ECC is coming up?mode_13h - Tuesday, February 14, 2023 - link
> Please expand on why ECC is coming up?This is sold as an industrial mini-PC. For something like that, reliability is key. Memory errors are one potential source of reliability problems, and ECC is an effective measure to compensate (short-term) and flag for replacement (long-term) any defective memory modules or boards.
The lore behind ECC is that it protects against cosmic rays, but I've only personally seen ECC errors that seem tied to flaky or failing hardware. It's worthwhile even for that purpose, alone.
TLindgren - Sunday, January 29, 2023 - link
It needs to be noted that SECDED over 512 bit is FAR less powerfull in handling errors than SECDED over 64-bit like regular ECC (or SECDED over 32-bit using DDR5 ECC sticks). They could have instead emulated the SECDED over each 64-bit chunk but then the extra reserved memory would have needed to be 8GB instead of 2GB, and the performance penalty likely would have been sigificantly worse.SECDED means it's guaranteed to correct one incorrect bit (SEC) and detect two incorrect bits (DED), no warranties for what happen with more incorrect bits but there's a decent statistical chance it'll detect them (but no chance it'll fix them).
Obviously getting two or even three+ faulty bits in the same "group" is far more likely over 512-bit compared to 64-bit, in fact it's my understanding that it'll likely happen most of the time given how memory sticks are constructed!
It's still useful because it'll detect a certain percentage of the multi-bit error so you will often? get told that you that you have faulty memory (except this doesn't seem to work) before things crash which means you know you need to fix the hardware, but the "correct bits" part is unlikely to save you because at least some of the time it'll get multiple wrong bits in the burst. I suspect they would have been better of with just giving up on correcting and aiming for "detect as many bit errors as we can" (probably 3-4 guaranteed bit detected with the 16-bit of extra data per 512bit they choose).
It's definitely better than no ECC *if* the software support gets improved a bit, but is in no way comparable to "real" ECC. OTOH, it's not priced as that either but it needs to be pointed out because some people will sell it as if it is.
ganeshts - Monday, January 30, 2023 - link
Taken standalone, you arguments are completely sound.However, in the bigger picture, you should note that newer memory technologies include link ECC to protect the high-speed communication link between the SoC and the external memory, AND, the DRAM DIMMs themselves implement transparent ECC for the stored data.
Overall, even mission-critical requirements like ASIL / ISO26262 (for automotive safety) can be met with the requisite FIT (failure-in-time) rate using SECDED protection for 512-bit blocks *assuming those other protection mechanisms are also in place*.
In-band ECC is also used on Tegra for such embedded applications [ https://twitter.com/never_released/status/13559704... ; I can't seem to dig up the original documentation, but remember this was heavily discussed when the Tegra feature was made public ].
ganeshts - Monday, January 30, 2023 - link
(Correction: DRAM DIMMs -> The memory chips)mode_13h - Tuesday, February 14, 2023 - link
> you should note that newer memory technologies include link ECC to protect the high-speed communication link between the SoC and the external memoryAre you saying the system you reviewed also supports traditional out-of-band ECC? Why wasn't that mentioned in the review? If not, then your point would seem to be moot.
I also don't see the point of using in-band ECC atop OOB ECC. Anything that OOB ECC can't correct doesn't seem like it's going to be correctable by in-band ECC.