On Thu, 26 May 2022 13:55:39 -0700 Wei Xu <weixugc@xxxxxxxxxx> wrote: > On Thu, May 26, 2022 at 12:39 AM Ying Huang <ying.huang@xxxxxxxxx> wrote: > > > > On Thu, 2022-05-26 at 00:08 -0700, Wei Xu wrote: > > > On Wed, May 25, 2022 at 11:55 PM Ying Huang <ying.huang@xxxxxxxxx> wrote: > > > > > > > > On Wed, 2022-05-25 at 20:53 -0700, Wei Xu wrote: > > > > > On Wed, May 25, 2022 at 6:10 PM Ying Huang <ying.huang@xxxxxxxxx> wrote: > > > > > > > > > > > > On Wed, 2022-05-25 at 08:36 -0700, Wei Xu wrote: > > > > > > > On Wed, May 25, 2022 at 2:03 AM Ying Huang <ying.huang@xxxxxxxxx> wrote: > > > > > > > > > > > > > > > > On Tue, 2022-05-24 at 22:32 -0700, Wei Xu wrote: > > > > > > > > > On Tue, May 24, 2022 at 1:24 AM Ying Huang <ying.huang@xxxxxxxxx> wrote: > > > > > > > > > > > > > > > > > > > > On Tue, 2022-05-24 at 00:04 -0700, Wei Xu wrote: > > > > > > > > > > > On Thu, May 19, 2022 at 8:06 PM Ying Huang <ying.huang@xxxxxxxxx> wrote: > > > > > > > > > > > > > > > > > > > > > > > > On Wed, 2022-05-18 at 00:09 -0700, Wei Xu wrote: > > > > > > > > > > > > > On Thu, May 12, 2022 at 8:00 AM Jonathan Cameron > > > > > > > > > > > > > <Jonathan.Cameron@xxxxxxxxxx> wrote: > > > > > > > > > > > > > > > > > > > > > > > > > > > > On Wed, 11 May 2022 23:22:11 -0700 > > > > > > > > > > > > > > Wei Xu <weixugc@xxxxxxxxxx> wrote: > > > > > > > > > > > > > > > The current kernel has the basic memory tiering support: Inactive > > > > > > > > > > > > > > > pages on a higher tier NUMA node can be migrated (demoted) to a lower > > > > > > > > > > > > > > > tier NUMA node to make room for new allocations on the higher tier > > > > > > > > > > > > > > > NUMA node. Frequently accessed pages on a lower tier NUMA node can be > > > > > > > > > > > > > > > migrated (promoted) to a higher tier NUMA node to improve the > > > > > > > > > > > > > > > performance. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > In the current kernel, memory tiers are defined implicitly via a > > > > > > > > > > > > > > > demotion path relationship between NUMA nodes, which is created during > > > > > > > > > > > > > > > the kernel initialization and updated when a NUMA node is hot-added or > > > > > > > > > > > > > > > hot-removed. The current implementation puts all nodes with CPU into > > > > > > > > > > > > > > > the top tier, and builds the tier hierarchy tier-by-tier by establishing > > > > > > > > > > > > > > > the per-node demotion targets based on the distances between nodes. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > This current memory tier kernel interface needs to be improved for > > > > > > > > > > > > > > > several important use cases: > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > * The current tier initialization code always initializes > > > > > > > > > > > > > > > each memory-only NUMA node into a lower tier. But a memory-only > > > > > > > > > > > > > > > NUMA node may have a high performance memory device (e.g. a DRAM > > > > > > > > > > > > > > > device attached via CXL.mem or a DRAM-backed memory-only node on > > > > > > > > > > > > > > > a virtual machine) and should be put into a higher tier. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > * The current tier hierarchy always puts CPU nodes into the top > > > > > > > > > > > > > > > tier. But on a system with HBM (e.g. GPU memory) devices, these > > > > > > > > > > > > > > > memory-only HBM NUMA nodes should be in the top tier, and DRAM nodes > > > > > > > > > > > > > > > with CPUs are better to be placed into the next lower tier. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > * Also because the current tier hierarchy always puts CPU nodes > > > > > > > > > > > > > > > into the top tier, when a CPU is hot-added (or hot-removed) and > > > > > > > > > > > > > > > triggers a memory node from CPU-less into a CPU node (or vice > > > > > > > > > > > > > > > versa), the memory tier hierarchy gets changed, even though no > > > > > > > > > > > > > > > memory node is added or removed. This can make the tier > > > > > > > > > > > > > > > hierarchy unstable and make it difficult to support tier-based > > > > > > > > > > > > > > > memory accounting. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > * A higher tier node can only be demoted to selected nodes on the > > > > > > > > > > > > > > > next lower tier as defined by the demotion path, not any other > > > > > > > > > > > > > > > node from any lower tier. This strict, hard-coded demotion order > > > > > > > > > > > > > > > does not work in all use cases (e.g. some use cases may want to > > > > > > > > > > > > > > > allow cross-socket demotion to another node in the same demotion > > > > > > > > > > > > > > > tier as a fallback when the preferred demotion node is out of > > > > > > > > > > > > > > > space), and has resulted in the feature request for an interface to > > > > > > > > > > > > > > > override the system-wide, per-node demotion order from the > > > > > > > > > > > > > > > userspace. This demotion order is also inconsistent with the page > > > > > > > > > > > > > > > allocation fallback order when all the nodes in a higher tier are > > > > > > > > > > > > > > > out of space: The page allocation can fall back to any node from > > > > > > > > > > > > > > > any lower tier, whereas the demotion order doesn't allow that. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > * There are no interfaces for the userspace to learn about the memory > > > > > > > > > > > > > > > tier hierarchy in order to optimize its memory allocations. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > I'd like to propose revised memory tier kernel interfaces based on > > > > > > > > > > > > > > > the discussions in the threads: > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > - https://lore.kernel.org/lkml/20220425201728.5kzm4seu7rep7ndr@offworld/T/ > > > > > > > > > > > > > > > - https://lore.kernel.org/linux-mm/20220426114300.00003ad8@xxxxxxxxxx/t/ > > > > > > > > > > > > > > > - https://lore.kernel.org/linux-mm/867bc216386eb6cbf54648f23e5825830f5b922e.camel@xxxxxxxxx/T/ > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > High-level Design Ideas > > > > > > > > > > > > > > > ======================= > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > * Define memory tiers explicitly, not implicitly. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > * Memory tiers are defined based on hardware capabilities of memory > > > > > > > > > > > > > > > nodes, not their relative node distances between each other. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > * The tier assignment of each node is independent from each other. > > > > > > > > > > > > > > > Moving a node from one tier to another tier doesn't affect the tier > > > > > > > > > > > > > > > assignment of any other node. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > * The node-tier association is stable. A node can be reassigned to a > > > > > > > > > > > > > > > different tier only under the specific conditions that don't block > > > > > > > > > > > > > > > future tier-based memory cgroup accounting. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > * A node can demote its pages to any nodes of any lower tiers. The > > > > > > > > > > > > > > > demotion target node selection follows the allocation fallback order > > > > > > > > > > > > > > > of the source node, which is built based on node distances. The > > > > > > > > > > > > > > > demotion targets are also restricted to only the nodes from the tiers > > > > > > > > > > > > > > > lower than the source node. We no longer need to maintain a separate > > > > > > > > > > > > > > > per-node demotion order (node_demotion[]). > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Hi Wei, > > > > > > > > > > > > > > > > > > > > > > > > > > > > This proposal looks good to me, though we'll be having fun > > > > > > > > > > > > > > white boarding topologies from our roadmaps for the next few days :) > > > > > > > > > > > > > > > > > > > > > > > > > > That's good to hear. > > > > > > > > > > > > > > > > > > > > > > > > > > > A few comments inline. It also seems likely to me that there is little > > > > > > > > > > > > > > benefit in starting with 3 tiers as the maximum. Seems unlikely the > > > > > > > > > > > > > > code will be substantially simpler for 3 than it would be for 4 or 5. > > > > > > > > > > > > > > I've drawn out one simple case that needs 4 to do sensible things. > > > > > > > > > > > > > > > > > > > > > > > > > > We can make the number of tiers a config option. 3 tiers are just what > > > > > > > > > > > > > the kernel can reasonably initialize when there isn't enough hardware > > > > > > > > > > > > > performance information from the firmware. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Sysfs Interfaces > > > > > > > > > > > > > > > ================ > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > * /sys/devices/system/memtier/memtierN/nodelist > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > where N = 0, 1, 2 (the kernel supports only 3 tiers for now). > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Format: node_list > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Read-only. When read, list the memory nodes in the specified tier. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Tier 0 is the highest tier, while tier 2 is the lowest tier. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > The absolute value of a tier id number has no specific meaning. > > > > > > > > > > > > > > > What matters is the relative order of the tier id numbers. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > When a memory tier has no nodes, the kernel can hide its memtier > > > > > > > > > > > > > > > sysfs files. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > * /sys/devices/system/node/nodeN/memtier > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > where N = 0, 1, ... > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Format: int or empty > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > When read, list the memory tier that the node belongs to. Its value > > > > > > > > > > > > > > > is empty for a CPU-only NUMA node. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > When written, the kernel moves the node into the specified memory > > > > > > > > > > > > > > > tier if the move is allowed. The tier assignment of all other nodes > > > > > > > > > > > > > > > are not affected. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Initially, we can make this interface read-only. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Kernel Representation > > > > > > > > > > > > > > > ===================== > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > * All memory tiering code is guarded by CONFIG_TIERED_MEMORY. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > * #define MAX_MEMORY_TIERS 3 > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Support 3 memory tiers for now. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > * #define MEMORY_DEFAULT_TIER 1 > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > The default tier that a memory node is assigned to. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > * nodemask_t memory_tiers[MAX_MEMORY_TIERS] > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Store memory nodes by tiers. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > * int node_tier_map[MAX_NUMNODES] > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Map a node to its tier. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > For each CPU-only node c, node_tier_map[c] = -1. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Memory Tier Initialization > > > > > > > > > > > > > > > ========================== > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > By default, all memory nodes are assigned to the default tier > > > > > > > > > > > > > > > (MEMORY_DEFAULT_TIER). > > > > > > > > > > > > > > > > > > > > > > > > > > > > This is tighter than it needs to be. In many cases we can easily > > > > > > > > > > > > > > establish if there is any possibility of CPU being hotplugged into > > > > > > > > > > > > > > a memory node. If it's CXL attached no way CPUs are going to be > > > > > > > > > > > > > > turning up their later :) If CPU HP into a given node can't happen > > > > > > > > > > > > > > we can be more flexible and I think that often results in better decisions. > > > > > > > > > > > > > > See example below, though obviously I could just use the userspace > > > > > > > > > > > > > > interface to fix that up anyway or have a CXL driver move it around > > > > > > > > > > > > > > if that's relevant. In some other cases I'm fairly sure we know in > > > > > > > > > > > > > > advance where CPUs can be added but I'd need to check all the > > > > > > > > > > > > > > relevant specs to be sure there aren't any corner cases. I 'think' > > > > > > > > > > > > > > for ARM for example we know where all possible CPUs can be hotplugged > > > > > > > > > > > > > > (constraint coming from the interrupt controller + the fact that only > > > > > > > > > > > > > > virtual CPU HP is defined). > > > > > > > > > > > > > > > > > > > > > > > > > > We may not always want to put a CXL-attached memory device into a > > > > > > > > > > > > > slower tier because even though CXL does add some additional latency, > > > > > > > > > > > > > both the memory device and CXL can still be very capable in > > > > > > > > > > > > > performance and may not be much slower (if any) than the on-board DRAM > > > > > > > > > > > > > (e.g. DRAM from a remote CPU socket). > > > > > > > > > > > > > > > > > > > > > > > > > > Also, the default tier here is just the initial tier assignment of > > > > > > > > > > > > > each node, which behaves as if there were no tiering. A tiering > > > > > > > > > > > > > kernel init function can certainly reassign the tier for each node if > > > > > > > > > > > > > it knows enough about the hardware performance for these nodes from > > > > > > > > > > > > > the firmware. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > A device driver can move up or down its memory nodes from the default > > > > > > > > > > > > > > > tier. For example, PMEM can move down its memory nodes below the > > > > > > > > > > > > > > > default tier, whereas GPU can move up its memory nodes above the > > > > > > > > > > > > > > > default tier. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > The kernel initialization code makes the decision on which exact tier > > > > > > > > > > > > > > > a memory node should be assigned to based on the requests from the > > > > > > > > > > > > > > > device drivers as well as the memory device hardware information > > > > > > > > > > > > > > > provided by the firmware. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Memory Tier Reassignment > > > > > > > > > > > > > > > ======================== > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > After a memory node is hot-removed, it can be hot-added back to a > > > > > > > > > > > > > > > different memory tier. This is useful for supporting dynamically > > > > > > > > > > > > > > > provisioned CXL.mem NUMA nodes, which may connect to different > > > > > > > > > > > > > > > memory devices across hot-plug events. Such tier changes should > > > > > > > > > > > > > > > be compatible with tier-based memory accounting. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > The userspace may also reassign an existing online memory node to a > > > > > > > > > > > > > > > different tier. However, this should only be allowed when no pages > > > > > > > > > > > > > > > are allocated from the memory node or when there are no non-root > > > > > > > > > > > > > > > memory cgroups (e.g. during the system boot). This restriction is > > > > > > > > > > > > > > > important for keeping memory tier hierarchy stable enough for > > > > > > > > > > > > > > > tier-based memory cgroup accounting. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Hot-adding/removing CPUs doesn't affect memory tier hierarchy. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Memory Allocation for Demotion > > > > > > > > > > > > > > > ============================== > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > To allocate a new page as the demotion target for a page, the kernel > > > > > > > > > > > > > > > calls the allocation function (__alloc_pages_nodemask) with the > > > > > > > > > > > > > > > source page node as the preferred node and the union of all lower > > > > > > > > > > > > > > > tier nodes as the allowed nodemask. The actual target node selection > > > > > > > > > > > > > > > then follows the allocation fallback order that the kernel has > > > > > > > > > > > > > > > already defined. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > The pseudo code looks like: > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > targets = NODE_MASK_NONE; > > > > > > > > > > > > > > > src_nid = page_to_nid(page); > > > > > > > > > > > > > > > src_tier = node_tier_map[src_nid]; > > > > > > > > > > > > > > > for (i = src_tier + 1; i < MAX_MEMORY_TIERS; i++) > > > > > > > > > > > > > > > nodes_or(targets, targets, memory_tiers[i]); > > > > > > > > > > > > > > > new_page = __alloc_pages_nodemask(gfp, order, src_nid, targets); > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > The memopolicy of cpuset, vma and owner task of the source page can > > > > > > > > > > > > > > > be set to refine the demotion target nodemask, e.g. to prevent > > > > > > > > > > > > > > > demotion or select a particular allowed node as the demotion target. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Memory Allocation for Promotion > > > > > > > > > > > > > > > =============================== > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > The page allocation for promotion is similar to demotion, except that (1) > > > > > > > > > > > > > > > the target nodemask uses the promotion tiers, (2) the preferred node can > > > > > > > > > > > > > > > be the accessing CPU node, not the source page node. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Examples > > > > > > > > > > > > > > > ======== > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > ... > > > > > > > > > > > > > > > > > > > > > > > > > > > > > * Example 3: > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Node 0 & 1 are DRAM nodes, Node 2 is a memory-only DRAM node. > > > > > > > > > > > > > > > > > > > > > > > > > > > > Node2 is drawn as pmem. > > > > > > > > > > > > > > > > > > > > > > > > > > Typo. Good catch. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > All nodes are in the same tier. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > 20 > > > > > > > > > > > > > > > Node 0 (DRAM) ---- Node 1 (DRAM) > > > > > > > > > > > > > > > \ / > > > > > > > > > > > > > > > \ 30 / 30 > > > > > > > > > > > > > > > \ / > > > > > > > > > > > > > > > Node 2 (PMEM) > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > node distances: > > > > > > > > > > > > > > > node 0 1 2 > > > > > > > > > > > > > > > 0 10 20 30 > > > > > > > > > > > > > > > 1 20 10 30 > > > > > > > > > > > > > > > 2 30 30 10 > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > $ cat /sys/devices/system/memtier/memtier*/nodelist > > > > > > > > > > > > > > > <empty> > > > > > > > > > > > > > > > 0-2 > > > > > > > > > > > > > > > <empty> > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > $ cat /sys/devices/system/node/node*/memtier > > > > > > > > > > > > > > > 1 > > > > > > > > > > > > > > > 1 > > > > > > > > > > > > > > > 1 > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Demotion fallback order: > > > > > > > > > > > > > > > node 0: empty > > > > > > > > > > > > > > > node 1: empty > > > > > > > > > > > > > > > node 2: empty > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > * Example 4: > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Node 0 is a DRAM node with CPU. > > > > > > > > > > > > > > > Node 1 is a PMEM node. > > > > > > > > > > > > > > > Node 2 is a GPU node. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > 50 > > > > > > > > > > > > > > > Node 0 (DRAM) ---- Node 2 (GPU) > > > > > > > > > > > > > > > \ / > > > > > > > > > > > > > > > \ 30 / 60 > > > > > > > > > > > > > > > \ / > > > > > > > > > > > > > > > Node 1 (PMEM) > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > node distances: > > > > > > > > > > > > > > > node 0 1 2 > > > > > > > > > > > > > > > 0 10 30 50 > > > > > > > > > > > > > > > 1 30 10 60 > > > > > > > > > > > > > > > 2 50 60 10 > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > $ cat /sys/devices/system/memtier/memtier*/nodelist > > > > > > > > > > > > > > > 2 > > > > > > > > > > > > > > > 0 > > > > > > > > > > > > > > > 1 > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > $ cat /sys/devices/system/node/node*/memtier > > > > > > > > > > > > > > > 1 > > > > > > > > > > > > > > > 2 > > > > > > > > > > > > > > > 0 > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Demotion fallback order: > > > > > > > > > > > > > > > node 0: 1 > > > > > > > > > > > > > > > node 1: empty > > > > > > > > > > > > > > > node 2: 0, 1 > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > * Example 5: > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Node 0 is a DRAM node with CPU. > > > > > > > > > > > > > > > Node 1 is a GPU node. > > > > > > > > > > > > > > > Node 2 is a PMEM node. > > > > > > > > > > > > > > > Node 3 is a large, slow DRAM node without CPU. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Node 2 (PMEM) ---- > > > > > > > > > > > > > > > / | \ > > > > > > > > > > > > > > > / | 30 \ 120 > > > > > > > > > > > > > > > | | 100 \ > > > > > > > > > > > > > > > | Node 0 (DRAM) ---- Node 1 (GPU) > > > > > > > > > > > > > > > \ \ / > > > > > > > > > > > > > > > \ \ 40 / 110 > > > > > > > > > > > > > > > 80 \ \ / > > > > > > > > > > > > > > > --- Node 3 (Slow DRAM) > > > > > > > > > > > > > > > > > > > > > > > > > > > > This is close but not quite what was intended for Hesham's > > > > > > > > > > > > > > example... (note we just checked that Hesham's original node0-1 > > > > > > > > > > > > > > timing didn't make any sense.). > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > This was inspired by Hesham's example. But I should have also included > > > > > > > > > > > > > the version that illustrates the need to skip a tier when demoting > > > > > > > > > > > > > from certain nodes. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > node distances: > > > > > > > > > > > > > > > node 0 1 2 3 > > > > > > > > > > > > > > > 0 10 100 30 40 > > > > > > > > > > > > > > > 1 100 10 120 110 > > > > > > > > > > > > > > > 2 30 120 10 80 > > > > > > > > > > > > > > > 3 40 110 80 10 > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > $ cat /sys/devices/system/memtier/memtier*/nodelist > > > > > > > > > > > > > > > 1 > > > > > > > > > > > > > > > 0,3 > > > > > > > > > > > > > > > 2 > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > $ cat /sys/devices/system/node/node*/memtier > > > > > > > > > > > > > > > 1 > > > > > > > > > > > > > > > 0 > > > > > > > > > > > > > > > 2 > > > > > > > > > > > > > > > 1 > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > Demotion fallback order: > > > > > > > > > > > > > > > node 0: 2 > > > > > > > > > > > > > > > node 1: 0, 3, 2 > > > > > > > > > > > > > > > node 2: empty > > > > > > > > > > > > > > > node 3: 2 > > > > > > > > > > > > > > > > > > > > > > > > > > > > This is close but not quite the same as the example > > > > > > > > > > > > > > Hesham gave (note the node timing 1 to 0 on in the table > > > > > > > > > > > > > > with that example didn't make sense). I added another > > > > > > > > > > > > > > level of switching to make the numbers more obviously > > > > > > > > > > > > > > different and show how critical it might be. > > > > > > > > > > > > > > > > > > > > > > > > > > > > * Example 6: > > > > > > > > > > > > > > > > > > > > > > > > > > > > Node 0 is a DRAM node with CPU. > > > > > > > > > > > > > > Node 1 is a GPU node. > > > > > > > > > > > > > > Node 2 is a PMEM node. > > > > > > > > > > > > > > Node 3 is an extremely large, DRAM node without CPU. > > > > > > > > > > > > > > (Key point here being that it probably never makes sense > > > > > > > > > > > > > > to demote to anywhere else from this memory). > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > I've redone the timings wrt to example 5. > > > > > > > > > > > > > > Basis for this is 0 and 2 are directly connected > > > > > > > > > > > > > > via controllers in an SoC. 1 and 3 are connected > > > > > > > > > > > > > > via a a common switch one switch down switch > > > > > > > > > > > > > > (each hop via this is 100) > > > > > > > > > > > > > > All drams cost 10 once you've reached correct node > > > > > > > > > > > > > > and pmem costs 30 from SoC. > > > > > > > > > > > > > > Numbers get too large as a result but meh, I'm making > > > > > > > > > > > > > > a point not providing real numbers :) > > > > > > > > > > > > > > > > > > > > > > > > > > > > PMEM Node 2 > > > > > > > > > > > > > > |(30) > > > > > > > > > > > > > > CPU + DRAM Node0 > > > > > > > > > > > > > > |(100) > > > > > > > > > > > > > > Switch 1 > > > > > > > > > > > > > > |(100) > > > > > > > > > > > > > > Switch 2 > > > > > > > > > > > > > > (100) | |(100) > > > > > > > > > > > > > > Node 1 GPU Node3 Large memory. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > With one level of s > > > > > > > > > > > > > > > > > > > > > > > > > > > > Node 2 (PMEM) ---- > > > > > > > > > > > > > > / | \ > > > > > > > > > > > > > > / | 30 \ 330 > > > > > > > > > > > > > > | | 310 \ > > > > > > > > > > > > > > | Node 0 (DRAM) ---- Node 1 (GPU) > > > > > > > > > > > > > > \ \ / > > > > > > > > > > > > > > \ \ 310 / 210 > > > > > > > > > > > > > > 330 \ \ / > > > > > > > > > > > > > > --- Node 3 (Extremely large DRAM) > > > > > > > > > > > > > > > > > > > > > > > > > > > > To my mind, we should potentially also take into account > > > > > > > > > > > > > > the fact that Node3 can be known to never contain CPUs > > > > > > > > > > > > > > (in at least some architectures we know where the CPUs > > > > > > > > > > > > > > might be added later, they can't just magically turn up > > > > > > > > > > > > > > anywhere in the topology). > > > > > > > > > > > > > > > > > > > > > > > > > > > > node distances: > > > > > > > > > > > > > > node 0 1 2 3 > > > > > > > > > > > > > > 0 10 310 30 310 > > > > > > > > > > > > > > 1 310 10 330 210 > > > > > > > > > > > > > > 2 30 330 10 330 > > > > > > > > > > > > > > 3 310 210 330 10 > > > > > > > > > > > > > > > > > > > > > > > > > > > > So, my ideal would treat node 3 different from other dram nodes > > > > > > > > > > > > > > as we know it can't have CPUs. Trying to come up with an > > > > > > > > > > > > > > always correct order for nodes 3 and 2 is tricky as to a certain > > > > > > > > > > > > > > extent depends on capacity. If node 2 was big enough to take > > > > > > > > > > > > > > any demotion from node 0 and still have lots of room then demoting > > > > > > > > > > > > > > there form node 3 would make sense and visa versa. > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > $ cat /sys/devices/system/memtier/memtier*/nodelist > > > > > > > > > > > > > > 1 > > > > > > > > > > > > > > 0 > > > > > > > > > > > > > > 2 > > > > > > > > > > > > > > 3 > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > $ cat /sys/devices/system/node/node*/memtier > > > > > > > > > > > > > > 1 > > > > > > > > > > > > > > 0 > > > > > > > > > > > > > > 2 > > > > > > > > > > > > > > 3 > > > > > > > > > > > > > > > > > > > > > > > > > > > > Demotion fallback order: > > > > > > > > > > > > > > node 0: 2, 3 > > > > > > > > > > > > > > node 1: 3, 0, 2 (key being we will almost always have less pressure on node 3) > > > > > > > > > > > > > > node 2: 3 > > > > > > > > > > > > > > node 3: empty > > > > > > > > > > > > > > > > > > > > > > > > > > > > or as Hesham just pointed out this can be done with 3 tiers > > > > > > > > > > > > > > because we can put the GPU and CPU in the same tier because > > > > > > > > > > > > > > their is little reason to demote from one to the other. > > > > > > > > > > > > > > > > > > > > > > > > > > Thank you for the example. It makes sense to me to have node 3 on its > > > > > > > > > > > > > own tier. We can have either 3 tiers or 4 tiers in total (assuming > > > > > > > > > > > > > that the max number of tiers is a config option). > > > > > > > > > > > > > > > > > > > > > > > > > > > We are also a bit worried about ABI backwards compatibility because > > > > > > > > > > > > > > of potential need to make more space in tiers lower in number than > > > > > > > > > > > > > > CPU attached DDR. I rather liked the negative proposal with > > > > > > > > > > > > > > default as 0 that Huang, Ying made. > > > > > > > > > > > > > > > > > > > > > > > > > > It is hard to have negative values as the device IDs. > > > > > > > > > > > > > > > > > > > > > > > > > > The current proposal equals the tier device ID to the tier hierarchy > > > > > > > > > > > > > level, which makes the interface simpler, but less flexible. How > > > > > > > > > > > > > about the following proposal (which decouples the tier device ID from > > > > > > > > > > > > > the tier level)? > > > > > > > > > > > > > > > > > > > > > > > > > > /sys/devices/system/memtier/memtierN/nodelist > > > > > > > > > > > > > /sys/devices/system/memtier/memtierN/rank > > > > > > > > > > > > > > > > > > > > > > > > > > Each memory tier N has two sysfs files: > > > > > > > > > > > > > - nodelist: the nodes that are in this tier > > > > > > > > > > > > > - rank: an opaque value that helps decide the level at which this tier > > > > > > > > > > > > > is in the tier hierarchy (smaller value means faster tier) > > > > > > > > > > > > > > > > > > > > > > > > > > The tier hierarchy is determined by "rank", not by the device id > > > > > > > > > > > > > number N from "memtierN". > > > > > > > > > > > > > > > > > > > > > > > > > > The absolute value of "rank" of a memtier doesn't necessarily carry > > > > > > > > > > > > > any meaning. Its value relative to other memtiers decides the level of > > > > > > > > > > > > > this memtier in the tier hierarchy. > > > > > > > > > > > > > > > > > > > > > > > > > > The CPU-attached DRAM nodes are always in memtier0 (the device ID), > > > > > > > > > > > > > but memtier0 may not always be the top-tier, e.g. its level can be 3 > > > > > > > > > > > > > in a 5-tier system. > > > > > > > > > > > > > > > > > > > > > > > > > > For the above example (example 6), we can have: > > > > > > > > > > > > > > > > > > > > > > > > > > $ ls /sys/devices/system/memtier > > > > > > > > > > > > > memtier0 > > > > > > > > > > > > > memtier1 > > > > > > > > > > > > > memtier2 > > > > > > > > > > > > > memtier128 > > > > > > > > > > > > > > > > > > > > > > > > > > $ cat /sys/devices/system/memtier/memtier*/rank > > > > > > > > > > > > > 50 > > > > > > > > > > > > > 60 > > > > > > > > > > > > > 70 > > > > > > > > > > > > > 10 > > > > > > > > > > > > > > > > > > > > > > > > I understand that the device ID cannot be negtive. So we have to use > > > > > > > > > > > > rank. Can we make it possible to allow "rank" to be negtive? > > > > > > > > > > > > > > > > > > > > > > It is possible to allow "rank" to be negative, though I think all > > > > > > > > > > > positive values should work equally well. > > > > > > > > > > > > > > > > > > > > > > > Another choice is to do some trick on device ID. For example, the CPU- > > > > > > > > > > > > attached DRAM node are always memtier100 (the device ID). Then we can > > > > > > > > > > > > have memtier99, memtier100, memtier101, memteri102, .... That's not > > > > > > > > > > > > perfect too. > > > > > > > > > > > > > > > > > > > > > > If we go with the device ID tricks, one approach is to use sub-device IDs: > > > > > > > > > > > > > > > > > > > > > > - There are 3 major tiers: tier0 (e.g. GPU), tier1 (e.g.DRAM) and > > > > > > > > > > > tier2 (e.g. PMEM). > > > > > > > > > > > > > > > > > > > > > > - Each major tier can have minor tiers, e.g. tier0.0, tier1.0, > > > > > > > > > > > tier1.1, tier2.0, tier2.1. > > > > > > > > > > > > > > > > > > > > > > The earlier 4-tier example can be represented as: > > > > > > > > > > > > > > > > > > > > > > memtier0.0 -> memtier1.0 -> memtier2.0 -> memtier2.1 > > > > > > > > > > > > > > > > > > > > > > We can also omit .0 so that the tiers are: > > > > > > > > > > > > > > > > > > > > > > memtier0 -> memtier1 -> memtier2 -> memtier2.1 > > > > > > > > > > > > > > > > > > > > > > This should be flexible enough to support multiple tiers while keeping > > > > > > > > > > > the tier IDs relatively stable. > > > > > > > > > > > > > > > > > > > > > > It is not as flexible as the rank approach. For example, to insert a > > > > > > > > > > > new tier between 2.0 and 2.1, we need to add a tier 2.2 and reassign > > > > > > > > > > > existing nodes to these 3 tiers. Using "rank", we can insert a new > > > > > > > > > > > tier and only move desired nodes into the new tier. > > > > > > > > > > > > > > > > > > > > > > What do you think? > > > > > > > > > > > > > > > > > > > > The rank approach looks better for. And if we stick with the device ID > > > > > > > > > > rule as follows, > > > > > > > > > > > > > > > > > > > > ... > > > > > > > > > > 255 GPU > > > > > > > > > > 0 DRAM > > > > > > > > > > 1 PMEM > > > > > > > > > > 2 > > > > > > > > > > ... > > > > > > > > > > > > > > > > > > > > 255 is -1 for "s8". > > > > > > > > > > > > > > > > > > > > The device ID should do most tricks at least now. The rank can provide > > > > > > > > > > more flexibility in the future. We can even go without rank in the > > > > > > > > > > first version, and introduce it when it's necessary. > > > > > > > > > > > > > > > > > > Given that the "rank" approach is generally favored, let's go with > > > > > > > > > that to avoid compatibility issues that may come from the switch of > > > > > > > > > device ID tricks to ranks. > > > > > > > > > > > > > > > > OK. Just to confirm. Does this mean that we will have fixed device ID, > > > > > > > > for example, > > > > > > > > > > > > > > > > GPU memtier255 > > > > > > > > DRAM (with CPU) memtier0 > > > > > > > > PMEM memtier1 > > > > > > > > > > > > > > > > When we add a new memtier, it can be memtier254, or memter2? The rank > > > > > > > > value will determine the real demotion order. > > > > > > > > > > > > > > With the rank approach, the device ID numbering should be flexible and > > > > > > > not mandated by the proposal. > > > > > > > > > > > > If so, the rank number will be fixed? For example, > > > > > > > > > > > > GPU 100 > > > > > > DRAM (with CPU) 200 > > > > > > PMEM 300 > > > > > > > > > > > > When we add a new memtier, its rank can be 50, 150, 250, or 400? > > > > > > > > > > > > If so, this makes me think why we don't just make this kind of rank the > > > > > > device ID? Or I missed something? > > > > > > > > > > > > Or, both device IDs and rank values are not fixed? Why do we need that > > > > > > kind of flexibility? Sorry, I may not undersand all requirements. > > > > > > > > > > Even though the proposal doesn't mandate a particular device ID > > > > > numbering, I expect that the device IDs will be relatively stable once > > > > > a kernel implementation is chosen. For example, it is likely that DRAM > > > > > nodes with CPUs will always be on memtier1, no matter how many tiers > > > > > are higher or lower than these nodes. > > > > > > > > > > We don't need to mandate a particular way to assign the rank values, > > > > > either. What matters is the relative order and some reasonable gap > > > > > between these values. > > > > > > > > > > The rank approach allows us to keep memtier device IDs relatively > > > > > stable even though we may change the tier ordering among them. Its > > > > > flexibility can have many other uses as well. For example, we can > > > > > insert a new memtier into the tier hierarchy for a new set of nodes > > > > > without affecting the node assignment of any existing memtier, > > > > > provided that there is enough gap in the rank values for the new > > > > > memtier. > > > > > > > > > > Using the rank value directly as the device ID has some disadvantages: > > > > > - It is kind of unconventional to number devices in this way. > > > > > - We cannot assign DRAM nodes with CPUs with a specific memtier device > > > > > ID (even though this is not mandated by the "rank" proposal, I expect > > > > > the device will likely always be memtier1 in practice). > > > > > - It is possible that we may eventually allow the rank value to be > > > > > modified as a way to adjust the tier ordering. We cannot do that > > > > > easily for device IDs. > > > > > > > > OK. I can understand that sometimes it's more natural to change the > > > > order of a set of nodes with same memory types (and data plane path) > > > > together instead of change that one by one for each node. > > > > > > > > It appears that the memtierX device becomes kind of memory types (with > > > > data plane path considered for latency/throughput too). We can assign a > > > > memory type for a node, and change the order between memory types. If > > > > so, we need to allow multiple memtiers have same rank value. > > > > > > Jonathan mentioned this feature that multiple memtiers share the same > > > rank as well. It can be a convenient feature to have. For > > > simplicity, it should be fine to leave out this feature initially. > > > > OK. What do you think about the concept of memory types? You have > > mentioned that in memtierX directory, we can put latency/throughput, > > etc. IMHO, these only make sense for one type of memory. And it's > > natural for all memory nodes onlined by a driver to be same memory type. > > I think this is not always true. For example, a dax kmem driver can > online both pmem and non-pmem dax devices as system memory. CXL Type 3 memory driver is also responsible for a memories of different types with very different characteristics. Would need to assign memory into at least a few different tiers - potentially many different ones. > > > That is, drivers (including firmware drivers) will register memory types > > and put nodes into it. Base on memory types, "rank" (related to for > > example latency) determined the real memory tiers. > > > > If you think it's a good idea, we can rename memtierX to memory_typeX. > > But memory type may be not a good name, DRAM in local memory controler > > and DRAM in remote CXL may have quite different performance metric. Or > > memory_class to avoid the possible confusion? > > Memory types (e.g. GPU, DRAM, PMEM, etc) can be useful information to > help initialize the memory tiers of NUMA nodes. But I think memory > type is not a substitute for memory tier. We still need to define > memory tiers on top of NUMA node groups based on memory types (for > example, some may want to group GPU and DRAM into the same tier, > others may want separate tiers for GPU/DRAM). It is simpler to keep > the sysfs interface to just memory tiers and implement memory types as > internal device attributes if needed. > > To avoid confusion, we can require that the rank value is unique for > each memtier device. This should make it clear that each memtier > device represents a distinct memory tier. I don't mind that for a first implementation, but can see advantage in flexibility of being able to have multiple tiers fuse by giving them the same rank value if we ever make rank writeable after creation. Given no userspace is going to rely on 'failure' to create ranks with the same value, the flexibility to make this change later without ABI compatibility problems is there. > We can still put > latency/throughput values into each memtierN directory. Such values > need to be specified as a range to better accommodate possibly varied > performance of the devices within the same memory tier. I'd postpone adding this sort of information to the tiers until we need it. Most of the info can be established by userspace anyway so why complicate this interface? If there are strong usecases for the info we can add it later. Thanks, Jonathan > > > Best Regards, > > Huang, Ying > > > > > > Best Regards, > > > > Huang, Ying > > > > > > > > > > > > > > > > > > I think you may need to send v3 to make sure everyone is at the same > > > > > > > > page. > > > > > > > > > > > > > > Will do it shortly. > > > > > > > > > > > > Good! Thanks! > > > > > > > > > > > > Best Regards, > > > > > > Huang, Ying > > > > > > > > > > > > > > Best Regards, > > > > > > > > Huang, Ying > > > > > > > > > > > > > > > > > > Best Regards, > > > > > > > > > > Huang, Ying > > > > > > > > > > > > > > > > > > > > > > > The tier order: memtier128 -> memtier0 -> memtier1 -> memtier2 > > > > > > > > > > > > > > > > > > > > > > > > > > $ cat /sys/devices/system/memtier/memtier*/nodelist > > > > > > > > > > > > > 0 > > > > > > > > > > > > > 2 > > > > > > > > > > > > > 3 > > > > > > > > > > > > > 1 > > > > > > > > > > > > > > > > > > > > > > > > > > $ ls -l /sys/devices/system/node/node*/memtier > > > > > > > > > > > > > /sys/devices/system/node/node0/memtier -> /sys/devices/system/memtier/memtier0 > > > > > > > > > > > > > /sys/devices/system/node/node1/memtier -> /sys/devices/system/memtier/memtier128 > > > > > > > > > > > > > /sys/devices/system/node/node2/memtier -> /sys/devices/system/memtier/memtier1 > > > > > > > > > > > > > /sys/devices/system/node/node3/memtier -> /sys/devices/system/memtier/memtier2 > > > > > > > > > > > > > > > > > > > > > > > > > > To override the memory tier of a node, we can use a new, write-only, > > > > > > > > > > > > > per-node interface file: > > > > > > > > > > > > > > > > > > > > > > > > > > /sys/devices/system/node/nodeN/set_memtier > > > > > > > > > > > > > > > > > > > > > > > > > > e.g. > > > > > > > > > > > > > > > > > > > > > > > > > > $ echo "memtier128" > sys/devices/system/node/node1/set_memtier > > > > > > > > > > > > > > > > > > > > > > > > I prefer the original proposal to make nodeX/memtier a normal file to > > > > > > > > > > > > hold memtier devicde ID instead of a link. > > > > > > > > > > > > > > > > > > > > > > OK. We don't have to use a symlink. > > > > > > > > > > > > > > > > > > > > > > > Best Regards, > > > > > > > > > > > > Huang, Ying > > > > > > > > > > > > > > > > > > > > > > > > > Any comments? > > > > > > > > > > > > > > > > > > > > > > > > > > > Jonathan > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > >
