UseNUMA membind Issue in openJDK

Gustavo Romero gromero at linux.vnet.ibm.com
Tue May 29 13:23:04 UTC 2018


Hi Swati,

On 05/29/2018 06:12 AM, Swati Sharma wrote:
> I have incorporated some changes suggested by you.
> 
> The use of struct bitmask's  maskp for checking 64 bit in single iteration
> is more optimized compared to numa_bitmask_isbitset()  as by using this we
> need to check each bit for  1024 times(SUSE case) and 64 times(Ubuntu Case).
> If its fine to iterate at initialization time then I can change.

Yes, I know, your version is more optimized. libnuma API should provide a
ready-made solution for that... but that's another story. I'm curious to know
what the time difference is on the worst case for both ways tho. Anyway, I
just would like to point out that, regardless performance, it's possible to
achieve the same result with current libnuma API.


> For the answer to your question:
> If it picks up node 16, not so bad, but what if it picks up node 0 or 1?
> It can be checked based on numa_distance instead of picking up the lgrps randomly.

That seems a good solution. You can do the checking very early, so
lgrp_spaces()->find() does not even fail (return -1), i.e. by changing the CPU to
node mapping on initialization (avoiding to change cas_allocate()). On that checking
both numa distance and if the node is bound (or not) would be considered to generate
the map.


Best regards,
Gustavo

> Thanks,
> Swati
> 
> 
> On Fri, May 25, 2018 at 4:54 AM, Gustavo Romero <gromero at linux.vnet.ibm.com <mailto:gromero at linux.vnet.ibm.com>> wrote:
> 
>     Hi Swati,
> 
> 
>     Thanks for CC:ing me. Sorry for the delay replying it, I had to reserve a few
>     specific machines before trying your patch :-)
> 
>     I think that UseNUMA's original task was to figure out the best binding
>     setup for the JVM automatically but I understand that it also has to be aware
>     that sometimes, for some (new) particular reasons, its binding task is
>     "modulated" by other external agents. Thanks for proposing a fix.
> 
>     I have just a question/concern on the proposal: how the JVM should behave if
>     CPUs are not bound in accordance to the bound memory nodes? For instance, what
>     happens if no '--cpunodebind' is passed and '--membind=0,1,16' is passed at
>     the same time on this numa topology:
> 
>     brianh at p215n12:~$ numactl -H
>     available: 4 nodes (0-1,16-17)
>     node 0 cpus: 0 1 2 3 8 9 10 11 16 17 18 19 24 25 26 27 32 33 34 35
>     node 0 size: 65342 MB
>     node 0 free: 56902 MB
>     node 1 cpus: 40 41 42 43 48 49 50 51 56 57 58 59 64 65 66 67 72 73 74 75
>     node 1 size: 65447 MB
>     node 1 free: 58322 MB
>     node 16 cpus: 80 81 82 83 88 89 90 91 96 97 98 99 104 105 106 107 112 113 114 115
>     node 16 size: 65448 MB
>     node 16 free: 63096 MB
>     node 17 cpus: 120 121 122 123 128 129 130 131 136 137 138 139 144 145 146 147 152 153 154 155
>     node 17 size: 65175 MB
>     node 17 free: 61522 MB
>     node distances:
>     node   0   1  16  17
>        0:  10  20  40  40
>        1:  20  10  40  40
>       16:  40  40  10  20
>       17:  40  40  20  10
> 
> 
>     In that case JVM will spawn threads that will run on all CPUs, including those
>     CPUs in numa node 17. Then once in
>     src/hotspot/share/gc/parallel/mutableNUMASpace.cpp, in cas_allocate():
> 
>       834 // This version is lock-free.
>       835 HeapWord* MutableNUMASpace::cas_allocate(size_t size) {
>       836   Thread* thr = Thread::current();
>       837   int lgrp_id = thr->lgrp_id();
>       838   if (lgrp_id == -1 || !os::numa_has_group_homing()) {
>       839     lgrp_id = os::numa_get_group_id();
>       840     thr->set_lgrp_id(lgrp_id);
>       841   }
> 
>     a newly created thread will try to be mapped to a numa node given your CPU ID.
>     So if that CPU is in numa node 17 it will then not find it in:
> 
>       843   int i = lgrp_spaces()->find(&lgrp_id, LGRPSpace::equals);
> 
>     and will fallback to a random map, picking up a random numa node among nodes
>     0, 1, and 16:
> 
>       846   if (i == -1) {
>       847     i = os::random() % lgrp_spaces()->length();
>       848   }
> 
>     If it picks up node 16, not so bad, but what if it picks up node 0 or 1?
> 
>     I see that if one binds mem but leaves CPU unbound one has to know exactly what
>     she/he is doing, because it can be likely suboptimal. On the other hand, letting
>     the node being picked up randomly when there are memory nodes bound but no CPUs
>     seems even more suboptimal in some scenarios. Thus, should the JVM deal with it?
> 
>     @Zhengyu, do you have any opinion on that?
> 
>     Please find a few nits / comments inline.
> 
>     Note that I'm not a (R)eviewer so you still need two official reviews.
> 
> 
>     Best regards,
>     Gustavo
> 
>     On 05/21/2018 01:44 PM, Swati Sharma wrote:
> 
>         ======================PATCH==============================
>         diff --git a/src/hotspot/os/linux/os_linux.cpp b/src/hotspot/os/linux/os_linux.cpp
>         --- a/src/hotspot/os/linux/os_linux.cpp
>         +++ b/src/hotspot/os/linux/os_linux.cpp
>         @@ -2832,14 +2832,42 @@
>              // Map all node ids in which is possible to allocate memory. Also nodes are
>              // not always consecutively available, i.e. available from 0 to the highest
>              // node number.
>         +  // If the nodes have been bound explicitly using numactl membind, then
>         +  // allocate memory from those nodes only.
> 
> 
>     I think ok to place that comment on the same existing line, like:
> 
>     -  // node number.
>     +  // node number. If the nodes have been bound explicitly using numactl membind,
>     +  // then allocate memory from these nodes only.
> 
> 
>              for (size_t node = 0; node <= highest_node_number; node++) {
>         -    if (Linux::isnode_in_configured_nodes(node)) {
>         +    if (Linux::isnode_in_bounded_nodes(node)) {
> 
>     ---------------------------------^ s/bounded/bound/
> 
> 
>                  ids[i++] = node;
>                }
>              }
>              return i;
>            }
>         +extern "C"  struct bitmask {
>         +  unsigned long size; /* number of bits in the map */
>         +  unsigned long *maskp;
>         +};
> 
> 
>     I think it's possible to move the function below to os_linux.hpp with its
>     friends and cope with the forward declaration of 'struct bitmask*` by using the
>     functions from numa API, notably numa_bitmask_nbytes() and
>     numa_bitmask_isbitset() only,  avoiding the member dereferecing issue and the
>     need to add the above struct explicitly.
> 
> 
>         +// Check if single memory node bound.
>         +// Returns true if single memory node bound.
> 
> 
>     I suggest a minuscule improvement, something like:
> 
>     +// Check if bound to only one numa node.
>     +// Returns true if bound to a single numa node, otherwise returns false.
> 
> 
>         +bool os::Linux::issingle_node_bound() {
> 
> 
>     What about s/issingle_node_bound/isbound_to_single_node/ ?
> 
> 
>         +  struct bitmask* bmp = _numa_get_membind != NULL ? _numa_get_membind() : NULL;
>         +  if(!(bmp != NULL && bmp->maskp != NULL)) return false;
> 
>                                -----^
>     Are you sure this checking is necessary? I think if numa_get_membind succeed
>     bmp->maskp is always != NULL.
> 
>     Indentation here is odd. No space before 'if' and return on the same line.
> 
>     I would try to avoid lines over 80 chars.
> 
> 
>         +  int issingle = 0;
>         +  // System can have more than 64 nodes so check in all the elements of
>         +  // unsigned long array
>         +  for (unsigned long i = 0; i < (bmp->size / (8 * sizeof(unsigned long))); i++) {
>         +    if (bmp->maskp[i] == 0) {
>         +      continue;
>         +    } else if ((bmp->maskp[i] & (bmp->maskp[i] - 1)) == 0) {
>         +      issingle++;
>         +    } else {
>         +      return false;
>         +    }
>         +  }
>         +  if (issingle == 1)
>         +    return true;
>         +  return false;
>         +}
>         +
> 
> 
>     As I mentioned, I think it could be moved to os_linux.hpp instead. Also, it
>     could be something like:
> 
>     +bool os::Linux::isbound_to_single_node(void) {
>     +  struct bitmask* bmp;
>     +  unsigned long mask; // a mask element in the mask array
>     +  unsigned long max_num_masks;
>     +  int single_node = 0;
>     +
>     +  if (_numa_get_membind != NULL) {
>     +    bmp = _numa_get_membind();
>     +  } else {
>     +    return false;
>     +  }
>     +
>     +  max_num_masks = bmp->size / (8 * sizeof(unsigned long));
>     +
>     +  for (mask = 0; mask < max_num_masks; mask++) {
>     +    if (bmp->maskp[mask] != 0) { // at least one numa node in the mask
>     +      if (bmp->maskp[mask] & (bmp->maskp[mask] - 1) == 0) {
>     +        single_node++; // a single numa node in the mask
>     +      } else {
>     +        return false;
>     +      }
>     +    }
>     +  }
>     +
>     +  if (single_node == 1) {
>     +    return true; // only a single mask with a single numa node
>     +  } else {
>     +    return false;
>     +  }
>     +}
> 
> 
>            bool os::get_page_info(char *start, page_info* info) {
>              return false;
>            }
>         @@ -2930,6 +2958,10 @@
>                                                           libnuma_dlsym(handle, "numa_bitmask_isbitset")));
>                  set_numa_distance(CAST_TO_FN_PTR(numa_distance_func_t,
>                                                   libnuma_dlsym(handle, "numa_distance")));
>         +      set_numa_set_membind(CAST_TO_FN_PTR(numa_set_membind_func_t,
>         +                                          libnuma_dlsym(handle, "numa_set_membind")));
>         +      set_numa_get_membind(CAST_TO_FN_PTR(numa_get_membind_func_t,
>         +                                          libnuma_v2_dlsym(handle, "numa_get_membind")));
>                  if (numa_available() != -1) {
>                    set_numa_all_nodes((unsigned long*)libnuma_dlsym(handle, "numa_all_nodes"));
>         @@ -3054,6 +3086,8 @@
>            os::Linux::numa_set_bind_policy_func_t os::Linux::_numa_set_bind_policy;
>            os::Linux::numa_bitmask_isbitset_func_t os::Linux::_numa_bitmask_isbitset;
>            os::Linux::numa_distance_func_t os::Linux::_numa_distance;
>         +os::Linux::numa_set_membind_func_t os::Linux::_numa_set_membind;
>         +os::Linux::numa_get_membind_func_t os::Linux::_numa_get_membind;
>            unsigned long* os::Linux::_numa_all_nodes;
>            struct bitmask* os::Linux::_numa_all_nodes_ptr;
>            struct bitmask* os::Linux::_numa_nodes_ptr;
>         @@ -4962,8 +4996,9 @@
>                if (!Linux::libnuma_init()) {
>                  UseNUMA = false;
>                } else {
>         -      if ((Linux::numa_max_node() < 1)) {
>         -        // There's only one node(they start from 0), disable NUMA.
>         +      if ((Linux::numa_max_node() < 1) || Linux::issingle_node_bound()) {
>         +        // If there's only one node(they start from 0) or if the process
>         +        // is bound explicitly to a single node using membind, disable NUMA.
>                    UseNUMA = false;
>                  }
>                }
>         diff --git a/src/hotspot/os/linux/os_linux.hpp b/src/hotspot/os/linux/os_linux.hpp
>         --- a/src/hotspot/os/linux/os_linux.hpp
>         +++ b/src/hotspot/os/linux/os_linux.hpp
>         @@ -228,6 +228,8 @@
>              typedef int (*numa_tonode_memory_func_t)(void *start, size_t size, int node);
>              typedef void (*numa_interleave_memory_func_t)(void *start, size_t size, unsigned long *nodemask);
>              typedef void (*numa_interleave_memory_v2_func_t)(void *start, size_t size, struct bitmask* mask);
>         +  typedef void (*numa_set_membind_func_t)(struct bitmask *mask);
>         +  typedef struct bitmask* (*numa_get_membind_func_t)(void);
>              typedef void (*numa_set_bind_policy_func_t)(int policy);
>              typedef int (*numa_bitmask_isbitset_func_t)(struct bitmask *bmp, unsigned int n);
>         @@ -244,6 +246,8 @@
>              static numa_set_bind_policy_func_t _numa_set_bind_policy;
>              static numa_bitmask_isbitset_func_t _numa_bitmask_isbitset;
>              static numa_distance_func_t _numa_distance;
>         +  static numa_set_membind_func_t _numa_set_membind;
>         +  static numa_get_membind_func_t _numa_get_membind;
>              static unsigned long* _numa_all_nodes;
>              static struct bitmask* _numa_all_nodes_ptr;
>              static struct bitmask* _numa_nodes_ptr;
>         @@ -259,6 +263,8 @@
>              static void set_numa_set_bind_policy(numa_set_bind_policy_func_t func) { _numa_set_bind_policy = func; }
>              static void set_numa_bitmask_isbitset(numa_bitmask_isbitset_func_t func) { _numa_bitmask_isbitset = func; }
>              static void set_numa_distance(numa_distance_func_t func) { _numa_distance = func; }
>         +  static void set_numa_set_membind(numa_set_membind_func_t func) { _numa_set_membind = func; }
>         +  static void set_numa_get_membind(numa_get_membind_func_t func) { _numa_get_membind = func; }
>              static void set_numa_all_nodes(unsigned long* ptr) { _numa_all_nodes = ptr; }
>              static void set_numa_all_nodes_ptr(struct bitmask **ptr) { _numa_all_nodes_ptr = (ptr == NULL ? NULL : *ptr); }
>              static void set_numa_nodes_ptr(struct bitmask **ptr) { _numa_nodes_ptr = (ptr == NULL ? NULL : *ptr); }
>         @@ -320,6 +326,15 @@
>                } else
>                  return 0;
>              }
>         +  // Check if node in bounded nodes
> 
> 
>     +  // Check if node is in bound node set. Maybe?
> 
> 
>         +  static bool isnode_in_bounded_nodes(int node) {
>         +    struct bitmask* bmp = _numa_get_membind != NULL ? _numa_get_membind() : NULL;
>         +    if (bmp != NULL && _numa_bitmask_isbitset != NULL && _numa_bitmask_isbitset(bmp, node)) {
>         +      return true;
>         +    } else
>         +      return false;
>         +  }
>         +  static bool issingle_node_bound();
> 
> 
>     Looks like it can be re-written like:
> 
>     +  static bool isnode_in_bound_nodes(int node) {
>     +    if (_numa_get_membind != NULL && _numa_bitmask_isbitset != NULL) {
>     +      return _numa_bitmask_isbitset(_numa_get_membind(), node);
>     +    } else {
>     +      return false;
>     +    }
>     +  }
> 
>     ?
> 
> 
>            };
>            #endif // OS_LINUX_VM_OS_LINUX_HPP
> 
> 
> 



More information about the hotspot-dev mailing list