SPD information
The SPD integrates the most important specifications of the kit on each module and can of course be read out. First of all, it is interesting that HWInfo recognizes the kit as DDR5-8200 – we’ll see why in a moment. Of course, the SKU number is also stored here and, in an exemplary fashion, we also find the manufacturer and revision of the memory modules here. Together with the capacity, this again gives us the image of SK Hynix 24 Gbit M-Die. The revision information is optional in the SPD. So at this point, praise to Corsair for including it.
The JEDEC specifications are given for DDR5-4800, as with the Corsair Vengeance 24 GB modules with Micron 24 Gbit RevB, which we last looked at. If we skip all the JEDEC timings, we come to the PMIC. As usual for consumer UDIMMs, this is of the PMIC5100 type. The manufacturer is Richtek Power and it is an “OC” type that can deliver more than 1.435 V VDD/VDDQ, which is particularly interesting for overclocking. Unfortunately, Corsair’s wording on their website is still somewhat misleading, which I had already noted in the last review of the Vengeance 24 GB modules. They refer to an “overclocked PMIC” when it is limited to 1.435 V. PMICs for more than 1.435 V, like the one in today’s kit, are referred to as “Extreme OC PMIC”. The SPD itself is of type SPD5118 and manufactured by Integrated Device Technology or Renesas.
Now we come to the reason why the kit is listed by HWInfo as DDR5-8200. The kit has 2 XMP profiles, the second of which is designed for DDR5-8200, and for HWInfo the second/last/fastest (?) profile “wins”. Why the kit is not advertised by Corsair as 8200 can only be speculated. The difficult XMP compatibility, which I’ll show you in a moment, could be part of the reason. In any case, the kit gives us two profiles to work with:
- DDR5-8000 with timings tCL 38, tRCD 38, tRP 38, tRAS 98, tRC 146, tWR 120, tRFC1 1638, tRFC2 879, tRFC_SB 759, CMD Rate 2N, at 1.4 V VDD, 1.4 V VDDQ, 1.8 V VPP and 1.2 V VDD2 (Memory Controller Voltage)
- DDR5-8200 with timings tCL 40, tRCD 52, tRP 52, tRAS 112, tRC 164, tWR 124, tRFC1 1685, tRFC2 904, tRFC_SB 781, CMD Rate 2N at 1.4 V VDD, 1.4 V VDDQ, 1.8 V VPP and 1.2 V VDD2 (Memory Controller Voltage)
It is immediately noticeable that the second profile uses significantly looser timings relative to the clock rate, but uses the same voltage values. At the end of the day, the performance should not really differ much. In my experience, the 1.2 V VDD2 in both profiles is utopically low and not sufficient for these clock rates with current Intel RAM controllers. Most motherboards ignore this value when loading the XMP profile and often set a more realistic value of e.g. 1.4 V – whether this is good or bad remains to be seen.
The supposedly supported topology of “2 DIMMs per channel”, which is stored in the first profile, is also somewhat strange. This is most likely a “typo”, as we are a long way from being able to run DDR5-8000 on a 4-DIMM mainboard with current CPUs/mainboards, let alone with full configuration on all RAM banks – but more on that later.
Heatsink test
As usual, we continue to test the cooling solution of the RAM modules and how much thermal potential is available for overclocking. Of course, this also implicitly tests whether a kit survives the stress test in XMP mode at all without becoming unstable – yes, that also happens! With the stress test Testmem5 v0.12 and the profile “Extreme1@Anta777” and without a slot distance between the modules, the highest possible thermal load is generated. Together with another temperature sensor on the testbench, the delta to the ambient temperature is determined, once passively cooled and once actively cooled with a 120 mm fan, which is located directly on the modules and blows downwards at 2000 rpm.
For the sake of clarity, the manufacturer and product names of the RAM kits have been abbreviated as follows:
- CDTR: Corsair Dominator Titanium RGB
- CDPR: Corsair Dominator Platinum RGB
- TGDR: Teamgroup DELTA RGB
- CVR: Corsair Vengeance RGB
- CV: Corsair Vengeance
- GSTZ5R: G.Skill Trident Z5 RGB
Here our kit is more in the midfield with 39.8 K without fan and 9.4 K with fan, which is still fully within the limits. Here you can already see the kit in a second expansion stage, namely without the removable RGB bar. This improves the temperatures somewhat, but not to the point. Of course, this may also be due to the fact that new RAM ICs are used in this kit, which we have not yet tested. If these also had a higher power consumption, this would be a potential explanation for the higher temperatures. The Corsair Vengeance kit with Micron 24 Gbit RevB ICs at the bottom of the diagram would confirm this theory.
As you may have noticed, I have given both variants an asterisk in the diagram. The reason for this is that the kit unfortunately cannot always survive the stress test without errors. The cause is unlikely to be the temperatures and the resulting instability, but something much more annoying – more on this in a few pages.
49 Antworten
Kommentar
Lade neue Kommentare
Urgestein
Urgestein
Veteran
Urgestein
Veteran
Veteran
Veteran
Urgestein
Urgestein
Urgestein
Urgestein
Mitglied
Urgestein
Mitglied
Urgestein
Urgestein
Veteran
Urgestein
Urgestein
Alle Kommentare lesen unter igor´sLAB Community →