mickulty

NICE score!

WOW

Great start! Looking forward to seeing more from it, fingers crossed for gold.

UD7, nice!

@I.nfraR.edAs far as I'm concerned it's the same principle, HOWEVER, there are existing scores in the base categories. That complicates things. Even if I made a full list, moving over all the relevant scores would still be a lot of work for some mod and I already owe leeg at least two beers. Also I know leeg sees it as a "trial" so messing with the existing database might not be a good idea. I think the practical answer for FX is to keep all 1T FX scores in the base category for now. That's not based on any long-term principle, just the fact that splitting them with FX would create a lot of mode work that doesn't exist with ADL. If the ultimate solution is something similar to phenom unlocking then it won't matter too much if there are 1c/CU results in the base rankings. BTW you can tell if you're in 1c/CU by looking at L2 cache in CPU-Z. Same L2 cache count as core count = 1 core per CU.

Hi folks, Since alder lake P-core categories have been added, and FX 1 core per CU categories have begun to be rolled out, I think clear guidance is needed on how to treat single-threaded CPU benchmarks (and validations, for that matter). Conventionally it's very common to disable cores for lightly threaded and singlethreaded work, but we're now in a situation where disabling idle cores can change the hardware category even though the work is being done by the same cores. Currently there seems to There are two approaches I can think of that make sense; Mandate that fully enabled scores go in the base hardware category, and scores with only P-cores enabled go in the P-core category. This would make pretty much the same OC of the same hardware, likely getting pretty much the same score, eligible for multiple categories. On the plus side, one could argue it would represent what the 1T capability of the hardware is in different configurations. Mandate that validations and 1T benches only go in the base category. This would probably be more enforcement effort in the short term, but in the long term the site could perhaps be written to handle it well (for example, removing P-core categories from the hardware list when uploading). On the plus side, you'd avoid having multiple rankings for the same hardware. Currently there seems to be some confusion, for example at least one bencher has what seems to be an 8P superpi 32M score in the base 12900K rankings while also having a score in the 8P rankings. I don't have a strong opinion, but my general inclination is that people have always disabled cores for lightly threaded benchmarks and it would be silly for it to just now start effecting the hardware category (therefore option #2 would be better). However I mostly just would like there to be a clear policy. It would be presumptive to start a poll myself, but perhaps one is called for?

Please keep reading;

Yes. Since apparently the change is going ahead for ADL, it would be a good idea to add one core per CU options for consistency across brands. I don't think going as far as listing the 8-core as a 4-core in all situations is necessary, but rather adding for example "FX-8350 (4P)" for one core per CU mode. Exception is the FX-4200 which is already one core per CU. This is especially relevant because AMD have indicated they're not pursuing big.little in the immediate future. CPU-Z shows the module count with the L2 and L1I cache configuration, so it is moderateable. It also does make a big difference for performance, see attached; So although the technical mechanism is not identical, disabling the "excess" cores on FX (and presumably A-series/athlon with dozer-derived cores) has much less performance impact than straight downcoring. AMD's so-called "CMT", like big.little, also sought to add more cores without the area cost of more full cores. Therefore they are far more similar to one another than either is to regular downcoring. ---------------------------------------------------------------------------------------- To those who, unlike roman, are against shaking things up for bulldozer-derivatives; I think a big part of why there's a bad taste in people's mouths is because this change is being pushed now. Not when mobile big.little appeared, not when lakefield appeared, but now, when it has an effect on an intel enthusiast launch. However, this big launch is the first time it's really been brought into question. The FX-8150 probably wouldn't have set GFPs in one core per CU mode anyway. So it is understandable. However since the change is being pushed I think it's important to the future of HWBOT to define things in a way that doesn't single out one vendor's approach. You can hardly blame people for seeing "HWBOT, supported by Intel", also seeing a change the benefits an Intel launch, and coming to a cynical conclusion. Therefore the opportunity to make it a general principle, not just "if AMD ever do exactly what Intel does", ought to be taken. This is why I previously suggested a general principle that could be applied that isn't just "if a company does what intel is doing..." Also the locked Intel SKUs need P-core categories too, obviously. Again, the current state where only WR-capable chips get the special ranking, would make someone think it's only so they can get WRs... that needs to be fixed. ---------------------------------------------------------------------------------------- Long-term moving to thread count might be a better idea. Otherwise, if someone decides to make monster cores with SMT3/SMT4, they might "unfairly" dominate. But that's a long process.

nice

No way I'm getting across the atlantic for this but grats on getting it organised! To clarify something you mentioned to me privately - just free entry, or free entry and free LN2?

*sweating*

Don't forget the CPU-Z validation!

Don't forget the CPU-Z validation!

Don't forget the CPU-Z validation!

Don't forget the CPU-Z validation!

This makes sense but there has to be a consistent standard applied to everyone, not just the new Intel chips. FX chips getting to play in 1 core per CU configurations would show it's not just a way to keep intel competitive for globals.

What rough area are you for collection on the chiller?

If globals are to be divided by core count then they should be divided by core count, that's that. Benchmarks should not be bent to how fast we feel a CPU should be. HWBOT has had big.little for a long time and always handled it this way, the difference is that now it affects a big boy. A mobile CPU is unlikely to be competitive in any global ranking anyway, and if it was that would be a bad thing for the site and community since they can't really be OC'd. Switching to thread count would be pretty reasonable from the perspective of "How do we compare this hardware?", but if that's done I want to know who's going to pick up the task of updating the database. To my understanding it's managed by volunteers at the moment. I don't think the site can reasonably embark on such a thing at the moment. Maybe in the future if the site has had the rewrite, had the bugs fixed that are waiting for the rewrite, and has paid stuff who can do it. Adding a separate hardware listing for disabled cores is not without precedent - after all, that's roughly how unlocking works with AM3. However, we need to think about when hardware qualifies for extra listings. I would propose; "When the average amount of execution resources per core is higher with some cores disabled, and it is possible to disable cores down to this configuration, an additional hardware listing should be added for the highest core count configuration that has the maximum amount of execution resources per core." Why this slightly awkward wording, you might ask? Well, all FX-based CPUs except the FX-4200 have execution resources shared between cores. And yet, unfairly, these CPUs have always had to compete with other chips that have the same core count. An FX-8350 has to compete with an i7-5960X! This proposal would allow an FX-8350 in one core per CU mode to compete fairly with an i7-4770K. It would still get roflstomped, but I think it's important on principle. This way there's a neutral wording, applied to everyone, rather than reference to a specific technology. It feels fairer. The effect is that a CPU with n strong cores gets to compete with other CPUs that also have n strong cores without the weak ones being relevant. The bit about it being possible matters because otherwise you'd get a load of random phone processors involved.

I'm shocked by the CPU restriction, when it said "same platform" I viewed that, as anyone would, to mean socket or maybe chipset. People talk about "all the CPUs on XXXXXX platform" all the time. Please can you announce all the stages ASAP so I know if I have to return my 5600G for a 5700G to be competitive, I was all in on this comp and bought a lot of hardware as well as encouraging others to compete, pretty unhappy to have what's effectively a new restriction dropped only a few days before.

TFW your top hwbot score roundup features your own scores

Strong stuff!

Thank you! For what it's worth all the Cardea SSDs currently on TEAMGROUP's site are; Cardea (in db already) Cardea Zero (in db already) Cardea II (NVMe 3.0 x4) Cardea Liquid (NVMe 3.0 x4) Cardea Zero Z440 (NVMe 4.0 x4) Cardea Zero Z330 (NVMe 3.0 x4) Cardea Zero Z340 (NVMe 3.0 x4) Cardea Ceramic C440 (NVMe 4.0 x4) Cardea IOPS (NVMe 3.0 x4) Cardea A440 (NVMe 4.0 x4) https://www.teamgroupinc.com/en/products/t-force/t-force-ssd#product-box-84 EDIT: As an afterthought: there is the "CARDEA II TUF Gaming Alliance" but that's just a Cardea II in a different colour

I have several SSDs to request the addition of. The SSD database is an extremely long way from complete, but these are disks that actually show up in the 1x AS SSD benchmark ranking at time of writing; Sabrent Rocket 4 Plus (NVMe - different to the Rocket NVMe 4.0) Teamgroup Cardea A440 (NVMe - at least I think that's what @mllrkllr88 is using? It's not the unqualified Cardea...) Samsung PM9A1 (NVMe) WD Blue SN550 (NVMe) HP EX900 (NVMe) Samsung PM981 (NVMe) Sandisk Ultra 3D (SATA SSD) SK Hynix BC501 (NVMe) FWIW I've attached my notes from going through the AS SSD 1x ranking - I'm planning to go through and report the scores that are matched to the wrong hardware once the new disks are added, seems unhelpful to do so before the correct hardware is there to be matched to. However, if strunk or antinomy wants to go through and correct them sooner, the links are in there. micks ssd vendetta.ods

good god